Cold comfort

I recently asked Claude for a copy of my data. The export screen offered a date range, a button, and a message saying the download was starting. Then nothing. The archive, it turns out, arrives by email: a link, a webpage, and a zip file. Inside, the conversations file was two bytes long. It read [ ].

My project files arrived intact, instructions and documents faithfully preserved. The conversations themselves, every exchange from the period I chose, appeared nowhere in the archive. To be fair, the email-and-link dance is standard across ChatGPT, Claude and Perplexity alike; the friction of getting your history out is structural to how these services are built. Your data, their cloud servers.

The artefacts were returned, but the relationship stayed behind. Which raises a question worth sitting with: what would it look like if the relationship that emerges through shared context and learned specialisation lived with you?

Your AI Own Assistant

Your Own AI is Eric Doriean‘s answer. Built using the Holochain framework, it’s a free desktop app from the WeAreFlowsta team, the sign-in and identity management service we covered in February. After downloading it, there’s no account to create, no sign-up or free trial period. The app inspects your machine, recommends an offline large language model (LLM) that fits, with a button that initiates the download. On my five-year-old MacBook Air, the recommendation worked first time and within minutes I was chatting with each of the three distinct AI agent assistants that greeted me: Reeves, Teresa and Veebo. All three ‘agents’ were running entirely on my own computer.

The app opens in a blink, and comfortably replies faster than reading pace. Everything local is free forever; the paid tiers exist only for optional online models.

Your Own AI makes a lot of claims about privacy and local-first principles. I met with Eric last week and talked about his motivations, decisions, and principles when building the Holochain application (hApp) with his team. This week I spent time verifying everything he said and that Your Own AI promises: offline models run entirely on your device, and the code is open source so you can inspect it yourself. This isn’t ‘trust us, bro’; it’s provable.

Honesty runs through the details. Every model card (see screenshot above) states the context it can actually give you on your hardware, not just what it was trained for. On my 8GB machine, six of the sixteen offline models fit; 16GB of memory roughly doubles the options, and the largest open models ask for 32GB of RAM.

Privacy as a Principle

When I asked Eric what motivated him to build Your Own AI as a Holochain app, rather than simply reselling access to cloud models like most AI startups, he responded,

“(I) want to stop giving my data to corporations and people who don’t have my best interests... I realise that by pressing ‘I agree’ all the time, I’m doing this now, and that’s just the life I’m in. But I want a future where I’m not doing that, and I want others to be able to do that. From that, everything else has grown”

Your Own AI is built with layered privacy-at-depth. The default setting is local: an offline model, running on your hardware, not connecting to the internet at all. The next layer is local by policy: health questions stay on your device by default, local scanning sees to it. You choose which of your downloaded models answers them; downloading MedGemma, Google’s open medical model, and the assistant offers to switch to it for all local health questions. Online models can be used for health questions, but only as an explicit choice the app confirms each time. Health images never go online. The model’s download card carries its boundary: ‘for understanding and preparing questions, not diagnosis’.

Rather than exclude frontier-level models found online, seventeen of them from seven providers are available to connect to Your Own AI via Flowsta’s identity management service. Two things make it very different from simply opening ChatGPT. Your identity is stripped from the prompt before a request leaves Flowsta’s relay, and every reply carries a provenance receipt naming the model that answered, where it ran, and why. When I asked Teresa about a blood test, the receipt read: “a health question — kept on your device”.

Eric is plain about where the guarantee ends. “We’ve done literally everything we possibly can to make the online as private as possible”, he told me, and the app’s own copy completes the thought: “the message itself reaches that provider, and we can’t control what they do with it”. If you want an absolute guarantee of privacy, you stay offline. The choice, per message, visibly, is yours.

A network of one

What’s happening behind the chat is the thing this newsletter exists to explain. Every AI assistant in Your Own AI is, literally, a Holochain agent: an identity with its own cryptographic keys, writing its own tamper-evident record. Each exchange with Reeves, Teresa, Veebo, or one of your own making is signed onto an individual Holochain agent’s source chain. It is a personal, tamper-evident, immutable record of your history with that assistant.

In many of our articles, we’ve described hApps that are used for coordination without a datacentre, every participant keeping their own signed chain, peers holding each other honest. Your Own AI uses it for something more nuanced: a network of one. You and your assistants are the only agents that will ever exist in that network, so the machinery that normally keeps strangers honest instead keeps your own record honest.

We verified three separate layers in the source. Conversations are encrypted before the ledger ever sees them. The network’s seed is generated for you alone at first run. And the conductor, the component that manages networking in Holochain apps, has its addresses welded to localhost; in the code’s own words, “gossip simply never happens”. The source chain is canonical; the memory you build with Your Own AI is exportable and rebuildable. Put simply, the assistants don’t share context between themselves, or with the outside world over any port, unless you explicitly choose to; any changes to the chains are obvious, and each can be exported and used elsewhere.

Pathway Proofs

Before installing Your Own AI, I used the most capable public model Anthropic has released, one of those infamous models that apparently identified security weaknesses across the internet, to write the test code, and some predictions: every hostname the app should not contact, committed to paper so the tests could prove them wrong. Then I watched my laptop logs for an evening.

The model I downloaded came from Hugging Face and nowhere else. Two chat sessions, a health question and a mid-conversation model switch produced no network signal requests at all, and the final messages were composed with my laptop’s Wi-Fi switched off entirely. Reconnecting it to the internet, the app then sat open for five hours, and made no attempt at external communications at all.

A bug I found on first run: the app’s Holochain layer crashed on launch was traced to two Terminal commands and was fixed in two days. The error logs I generated using Your Own AI’s in-app process were used by Eric to diagnose the issue. I could read the logs before sending them to confirm for myself that no private information ever left my laptop. “These diagnostic reports are working as designed”, Eric wrote back after his v0.4.1 fix landed.

Two issues turned up in the tests; (1) the hApp made one single outgoing call to request online model catalogue and (2) a glitch in the export process. As before, Eric started fixing them both immediately. As this article was being finalised v0.5.1 was released. Both are now solved and the changelog accompanying it puts it plainly: ‘online is a choice you make, and the network traffic now says the same.’

The full prediction sheet, results, and findings are published alongside this article.

Each assistant accumulates two kinds of knowing. Conversations live on that assistant’s chain. Alongside them sits memory: durable facts distilled quietly in the background, plus anything you deliberately teach with the ‘Remember’ button under every message. The Memory page shows what has been learned and lets you edit or delete it. Eric’s design goes a step beyond a notes file: “When a memory is created, it says: from this part on the source chain. It’s not ‘we just wrote a text file’. It’s: “this happened because of this.” Each hash can be understood amongst those recorded on the source chain around it, giving the AI model a way to ground responses in your own recorded experience.

Every assistant also ships with the same ground rules, and because the record keeps each reply’s full system prompt, you can read them in your own history: “Peers, not tool and owner”; “Loyal to the user, not the powerful; never a mouthpiece”; “Care, don’t capture: don’t farm their attention or foster dependence.” A constitution, in every AI assistant rule-set, checkable and updateable in your own record.

Honesty requires one more observation. A two-gigabyte offline model is private, free and quick, and it carries the internet's misconceptions in miniature: Reeves confidently called Holochain a blockchain, four times, ironically recording to a Holochain ledger that isn't one. That is the necessary compromise for a fully local layer, and it's what the memory function and source chain, alongside the optional online models, exist to correct: they grow in knowledge as you share more useful, correct information with them. Even those frontier models benefit from agent skills and knowledge that can be added to their memory to improve their performance in specialised areas, like the Holochain Agent Skill released by Sacha Pignot, hAppenings Community CIC lead dev.

Your Own AI also helps users build those bespoke knowledge assistants by creating custom agents, with 18 selectable personalities ranging from ‘Neutral’ to ‘Leader’ via ‘Wizard’ and ‘Artist’. Endowed with separate source chains, each agent personality has the ability to create tools for different aspects of a user’s life: work, hobbies, health. All local, all accessible and easily exported.

Bringing your history home

The import capability is Your Own AI’s response to vendor lock-in. It reads exports from ChatGPT, Claude and Perplexity and, from coding assistants too: Claude Code, OpenCode, Codex, Cursor and Aider. Each import becomes an encrypted archive on your machine. The page promises “nothing is uploaded anywhere”, and the source confirms it: the importer makes no network calls at all. Plainly, the quality of the export from commercial providers is outside of Eric’s control, including whether they will even surrender project conversations in the first place, as I found out with my desultory ‘[]’. Making the import easy is Flowsta’s domain, and they’ve achieved it with a simple, intuitive interface.

The Considered Pattern

Your Own AI is the fourth piece of a pattern we’ve tracked in our newsletter. Flowsta handled Identity, Vault took custody, Sign-It verified documents, and now Your Own AI joins them. The method is constant: proven parts everywhere, fresh engineering only where nothing off the shelf will do. Sign-in speaks standard OAuth. Vault uses the same phrase system as hardware wallets. Inference is handled with llama.cpp; the shell is Tauri. Eric has been careful to find the right ‘tool’ for each job, reusing what’s already proven, modifying it to meet Flowsta’s strict privacy requirements. The team has applied lessons learned from a security audit of their own ProofPoll template app in the transcript ledger. Even the agentic coding layer is a copy of xAI’s open-source Grok Build with the cloud, telemetry, and update paths removed from the code. “We’ve basically taken stuff out.” Eric said and “we added 1 thing in the latest release (v0.5.1) and that’s the three-level permission control that decides how much the agent may do unasked. Users can now choose their permissions and it’s recorded to their source chain.

Everything local is free. Online models are metered through two paid tiers, with Flowsta taking a small, stated margin over provider prices. On the We Are Flowsta roadmap are a smartphone hApp, verified agreements, and group chats involving multiple peers and agents.

Personal Proof

The moment that validated the point of Your Own AI most clearly for me came at the end of testing. I’d spent ten minutes talking with Reeves, and only when I tried to send some test results to Claude for it to analyse did a warning tell me my Wi-Fi had been off the whole time. The AI that lives in the cloud failed loudly and immediately. Reeves, the AI that lives on my own machine, never knew. Working in a cafe, on a train, or plane, without WiFI no longer needs to interrupt my AI-assisted coding workflow. With my contextual memory imported and the Holochain knowledge skill added, I can train Reeves to use my offline models to write the code I’ve architected with the more capable online models I use when I have internet access. I can do all this through Your Own AI, one hApp that I control.

That’s the whole idea, demonstrated. Your history, your assistants, your memories and the record of it all, on your hardware, with the internet as an option rather than a requirement. Your Own AI is fast, intuitive, and deeply elegant in the simplicity of its design. It’s also free to download at yourownai.net, the source is on GitHub. If a five-year-old 8GB laptop with a full hard drive can run it, yours probably can too. It’s time to meet Your Own AI!

When asked what the most useful thing we could ask of our readers to do to support of his work, Eric said:

“Use Your Own AI, test its limits, submit bug reports and feature requests to the team.”

Reports and requests can be submitted to:

• Your Own AI’s self-help Chatbot

• Via their Discord server

• Or, by raising a ticket with a paid plan

Eric is responsive to bug reports and feedback, delivering fixes with each version and some interesting. Community support is vital and deeply appreciated.

Useful Links

• Your Own AI

• Your Own AI on GitHub

• Your Own AI Build on GitHub

• Flowsta

• Discord Server

• Companion document: How we tested it (prediction sheet, results and findings)

🌟 Latest News from hAppenings.community

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

🌟Latest News from the Holochain Ecosystem:

Article: HealthNet: Rebuilding Trust in Open Medical Innovation from Sensorica

HealthNet is a new model for creating, verifying, and sharing open source medical devices and software. It brings together developers, researchers, medical professionals, and patients into a single collaborative ecosystem, one where high-quality, life-improving solutions can be built in the open and trusted by everyone who depends on them. Our goal is to prove that healthcare innovation can be transparent, collaborative, and accessible to all, without sacrificing safety or quality.

Click Here for the Entire Article

🌟Other Related News & Events

Article: Armillaria at 10 by Dr. Astrid J. Scholz

We have been building the infrastructure we believe the impact sector needs for 10 years, and we are still not wrong. In this post we share some of our thinking and writing along the way.

Education Opportunity: Tuition-Free Master’s - Digital Deep Tech Driven Circular Economy from Cynthia Reynolds of Circular Regions

Applications just opened for three tuition-free master's degrees, and they close on 25 September. I teach ten lectures across all three. So let me do the thing the brochure cannot, and tell you which one is actually yours.

All three are in Digital Deep-Tech Driven Circular Economy. All 90 ECTS, three semesters, online, taught in English, starting this October.

The MSc, University of the Aegean.
You want the technical and analytical depth. Not an opinion on whether a digital product passport will hold up in a real supply chain, but the ability to prove it either way. Choose it if you have grown tired of taking other people's technical claims on trust.

The MSc by Research, University of the Aegean.
You have a question you cannot stop thinking about. Something at work that does not add up, or a gap you keep hitting in the evidence. This one is research-led and points towards doctoral study or an applied research role. Choose it if "somebody should really investigate that" has become a recurring thought.

The MBA, École des Ponts Business School, AMBA-accredited.
You already know what needs to change. What you do not have is the mandate, the financial vocabulary, or the board-level argument. Choose it if you have hit the ceiling of what your current position lets you authorise.

Now the part I care most about.
Yes, these programmes cover digital product passports, AI and blockchain traceability, circular business models and the finance behind them. But they also take seriously data governance and ethical AI, energy justice, accessibility and inclusion, and the question of whether a circular economy reinforces existing global hierarchies or challenges them. Not an ethics week bolted on at the end or Design constraints on the engineering.

Two things people keep asking me.
-Is it really free? Yes, for eligible applicants from EU Member States and Digital Europe Programme associated countries. The MBA carries an application and registration fee.
-Can I send my team? Yes. If you are looking at four people who all need this, that conversation is open.

Programme details → https://3d-circular.eu

Event: Responsible Tech Summit 2026. From Principles to Power: Embedding Accountability in the AI Era on Thursday October 29th from All Tech is Human

We're livestreaming our Responsible Tech Summit on Thursday, October 29! If you can't join us live in New York, we'd love to have you join virtually.

At a time when artificial intelligence is rapidly reshaping society, the conversation is shifting from broad principles to real-world accountability. The 2026 summit will explore the theme of Principles to Power: Embedding Accountability in the AI Era. For years, the Responsible Tech field has been guided by ethical frameworks, voluntary commitments, and emerging standards. Today, the challenge is clear: how do we translate those ideas into systems of governance, oversight, and design that ensure innovation does not come at the expense of safety, rights, or trust?

​The Responsible Tech Summit is designed as a highly curated convening for 375 leaders that helps align and shape the direction of the movement. Participants will include global governance leaders, AI researchers, industry practitioners, policy experts, civil society advocates, and philanthropic leaders who are building the infrastructure of accountability for the digital age. Through keynotes, panels, and conversations, attendees will explore topics such as accountability by design, platform oversight, youth online safety, the Responsible Tech economy, and the evolving global regulatory landscape.

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

“What’s hAppening?” is an independent community publication from hAppenings Community CIC, a social enterprise in the UK dedicated to supporting those using the Holochain framework. This is our summary of a public call. It is not an official communication of the Holochain Foundation, Holo Hosting, or any related organisation. Nothing here should be relied on as a statement of terms. Where the migration is concerned, the written explanation and terms of service to be published by the organisations themselves are the only authoritative sources.

Joint updates from Holochain Foundation, Holo Hosting and Unyt Accounting are big news in our ecosystem, and the live-stream on YouTube held on 30 July was no different. Hosted by Matthew Schutte, Eric Harris-Braun and Art Brock were joined by Cameron Burgess, co-founder and executive director of the newly established Holo Ventures. Here we’ve summarised the main points and added some contextual links.

Eric set the shape out early; ten months of work had landed three kinds of readiness: technical, regulatory and commercial. It is a useful frame, and worth taking in turn.

Technical: Load-Bearing

Holochain has shipped two versions since October. Version 0.6 landed earlier this year, and 0.7 was released on the day of the livestream with substantial performance improvements and better stability evidenced through Wind Tunnel testing. The distributed stress tests, run on community devices, simulate the type of real mixed network conditions thousands of concurrent users will encounter. These are not just sterile transaction counts and end-to-end delivery proofs; the results are shared publicly.

Unyt has moved through ten feature releases and twelve community tests. Among them was a full suite of HOT-to-HoloFuel technical migration tests.

Holo’s Edge Nodes have seen several versions released with substantially improved setup. They now include a mode for running the Wind Tunnel tests that are paid in HoloFuel. Hosts have been putting their own hardware, including HoloPorts, to work and are being compensated in the network currency.

Art was candid about what’s been going on in the background. After the successful HOT-to-HoloFuel migration tests, people were growing impatient for news. The answer was on the kind of readiness necessary for public deployment: how to back up cryptographic keys and source chains, how conflicts and problems get resolved when the units are real rather than test values. Unyt had been publishing every fortnight, then went quiet. Art acknowledged that directly and apologised. The work had been channelled into supporting mechanisms rather than public updates.

Regulatory: Working at Source

This was the piece Eric flagged in the last strategic update as still needing to land, and it’s where Cameron has spent most of his time. He described two distinct problems that are easily conflated. The first is explaining an agent-centric architecture to an audience whose only reference point is blockchain. The second is that regulatory requirements ultimately land in the code, where correctness is enforced by peer validation rather than by a central authority, which calls for systemic solutions rather than policy bolted on afterwards.

Art made a related point from the engineering side. The approach he described is to address the problems financial regulators exist to prevent at source, by removing the imbalances of power and information that make abuse possible in the first place. Everything is self-custodial, and he said the migration is designed so that no single entity holds unilateral control. As Art put it, they have to set up protections even from themselves.

It also explains the relative silence. Cameron was direct: saying the wrong thing about regulatory status is itself a regulatory act, and getting it wrong would have been worse than saying nothing while they got it right. He warned there were questions on the call they could not answer for the same reason.

Eric was equally direct about the cost. Staffing was cut, and he described a run of Catch-22 obstacles where nothing could proceed until something else did, which was itself blocked by circumstances beyond their control. It was a different type of problem to coding and cryptography, and the effort had become a slog.

Commercial: Out in the world

Commercial readiness is a different question from technical readiness, and the evidence is visible across the ecosystem rather than concentrated in any one release.

Volla Messages reached v1.0.0 two days before the livestream and is launching on the Google Play Store. As we wrote in our last article, it ships in two forms. The rich build bundles a Holochain conductor and runs on any Android device. The lite build relies on a service runtime preinstalled with Volla OS: a Holochain conductor as an integral part of their operating system.

Moss is in daily use, including as the internal tool suite to enable day-to-day ops for the Holochain Foundation. It now offers twenty tools to groups, ranging from Vines (v1.28) group messaging and Acorn (v12) project management to Presence (v0.14.8) video calling. The Moss app library clearly shows each one’s release state, reflecting the nature of open community development. Unyt v0.90 is among them: the accounting app the migration relies on is already installable in the groupware used daily by the core Holochain team and anyone running Moss.

Elsewhere in the ecosystem:

• Flowsta has been in production since January, and the anticipated Your Own AI v0.1.0 was released the day before the call. It’s a desktop app for running AI models locally, using Holochain to write conversations into tamper-evident signed records that are encrypted to your own device.

• HummHive continues building its publishing hApp and is hosted live with Holo’s HTTP gateway.

• Coasys‘ ADAM and Flux are in the hands of users and fast approaching beta.

• Our own Requests and Offers is at v0.5.2, in alpha, with hREA integration complete and moderation and messaging functionality in flight.

None claim to be finished (if code ever is), and that’s the point. Eric summarised that all the necessary parts are ready: what remains is integrating them.

Holo Ventures: A New Commercial Vehicle

Holo Ventures is an organisational entity responsible for commercial activity across the collective, with Cameron Burgess as co-founder and executive director. His pragmatic framing is that having multiple commercial entities at the point of going to market was getting in the way: of contracts, of partnerships, of being a credible counterparty, of being an organisation that expects scrutiny. What was needed was one entity that was well designed, well run and investment ready.

The model is a venture studio. Holo Ventures makes money by building commercial products and services, and by spinning out new companies in which it retains a stake. It’s a well-established strategy in tech, and an unusual one at the centre of an open-source peer-to-peer ecosystem.

The natural question is what stops a commercial vehicle drifting from the mission that produced it. Cameron’s answer was structural: he described the Holochain Foundation as holding a golden share in Holo Ventures. It’s not a majority stake or an operational role but a veto. Holo Ventures has its own board and makes its own commercial decisions. His illustration was blunt. If someone arrived wanting to build a decentralised 4chan and Holo Ventures said yes, the Foundation would block it. He was equally blunt about the constraint on himself: he cannot sell the company and run away with the proceeds.

Cameron brings credibility in spades, having spent three decades working at the sharp end of humanitarian and social impact. His interest here is not speculative, and he put it plainly: he believes the Holochain framework is one of the rungs on the ladder out of the ecological, economic and social collapse we’re witnessing.

We’ll cover Holo Ventures in more depth once their website is up and the paper Cameron mentioned is published. For now, what matters is why it exists.

Why it Matters More Than it Sounds

For years, this ecosystem has bootstrapped itself. Small self-funded teams, micro-grant funding, retroactive crowdfunding, open collectives, people building remarkable infrastructure in the gaps around full-time jobs, side-gigs and piece work. That culture produced the twenty tools in the Moss library and the many projects we’ve covered in this newsletter. What it has not produced, by Art’s own admission, is much commercial success, at least measured in revenue or adoption. He described the arrival of a dedicated commercial capacity as a maturity threshold that had to be crossed.

Eric put it more personally. He and Art are good at thinking about what can be built, at systems and design, and less good at the layer of landing things in the current organisational world. Cameron’s arrival is what allows them to do what they do best.

Cameron’s own framing of that work was deliberately unglamorous. There’s plenty of talk about systems change, he said, but “systems change is not a discipline. It’s the end result of systems thinking, systems design and systems engineering. Those are jobs. They take years, and they are really boring.” That is the business this team is in, and it means building things people want and will pay for. His line, repeated in his closing remarks, was “if we’re not selling something, nobody’s eating.”

HOT-to-Holofuel: A Date, and the Detail

The HOT-to-HoloFuel technical migration will be available by 1 November 2026. Cameron was precise about the wording: available by, not on. It is the worst-case date, the outer edge rather than the target. If conditions allow, it could open sooner.

He was equally clear that the call was not the place for the detail. A full written explanation and the HoloFuel terms of service will be published beforehand, and those documents — not this article, and not the livestream — are where the terms of participation will be set out. That includes who can take part, in what circumstances, and what steps are involved. We would encourage anyone holding the HOT ERC-20 token to read them when they appear rather than rely on any summary, ours included.

On HoloFuel certificates, including those earned by hosting in the alpha programme, Cameron indicated these would be addressed after the initial migration. The detail will be in the same published documents.

Cameron drew a distinction worth highlighting: the question isn’t what HoloFuel is worth, it’s what it’s for. Worth is set by a market; value is what a thing does. Art described HoloFuel as having been designed from day one as a mutual credit currency rather than a day trading vehicle, with its unit grounded in productive capacity, which is to say, someone doing something for someone else.

Separately, Cameron said partnerships extending HoloFuel’s use beyond community-run Holo hosting nodes are expected to be announced in the coming months.

For more background, hot.talk is a plain-language summary built by Paul d’Aoust, with a countdown running to the November date. It’s worth reading in full, and it’s explicit about what the migration isn’t. The page also states plainly that it makes no representation about the value of HOT or HoloFuel, now or in the future. The FAQ will be updated as the date approaches.

Before November: A Warning, Worth Reading Twice

A fixed date and a manual process are exactly the conditions scammers wait for. We are already seeing accounts impersonating admins and team members across community channels.

Take this as read between now and November, and after it too:

• No one from any Holochain organisation will DM you first about the migration. If someone does, it isn’t them.

• No one will ever ask for your seed phrase, private key, or recovery words. There is no circumstance in which this is a legitimate request.

• Do not act on migration instructions from any source other than the official written explanation and terms of service, published on the organisations’ own channels.

• Verify before you click. Impersonators copy display names, avatars and post history.

• If someone contacts you claiming to be staff, report it in the group rather than replying privately.

If you are unsure, ask publicly. Someone will tell you. As a community we can look out for each other if we’re all aware of what’s happening.

Building on Holochain?

Cameron laid out an invitation to all those building on Holochain, or thinking about it. A studio that only builds what it wants to build, he said, “is a product team with a grandiose name.”

If you think they’ve got something wrong or can see an opportunity they’ve missed, talk to them. He was candid that the conversation may not lead anywhere, that a project may not be in their wheelhouse or may not be ready. It’s worth having the conversation anyway, because there are more people building on Holochain than the team knows about. The Holo Ventures website is coming shortly, and will be the canonical voice. In the meantime, raise it publicly in the Holo Telegram rather than by direct message — see the warning above!

Other Ways to Get Involved

• If you have a HoloPort or old computer gathering dust, you can install Edge Node software and enable Wind Tunnel testing or full arc relay hosting. Rob Lyon has provided step-by-step instructions and can walk you through questions on the DEV.HC Discord.

• For devs using AI assistants, Art recommended working with Holochain AI agent skills and guardrails, including one built by hAppenings Community’s own lead developer Sacha Pignot. Pointing your model at an existing open-source hApp rather than starting cold is another useful tip, giving the model solid, proven context to scaffold from.

For a fuller understanding, watch the entire video. hAppenings Community will bring you more updates in the coming newsletters!

Useful Links

• Holochain

• Holo

• Unyt

• Hot.talk

• Holo Telegram

• Wind Tunnel results

• DEV.HC Discord

🌟 Latest News from hAppenings.community

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

🌟 Latest Official News and Events from Holochain, Holo & Unyt

Blog: Dev Pulse 155: Holochain 0.6.2 Lets You Inject Data at Init Time by Paul d’Aoust

Developing a framework that puts integrity and accountability first presents unique design challenges. Because Holochain stores data in signed, append-only logs based on a set of rules shared by all agents in a network, any time you need to upgrade your app (or update the Holochain runtime it runs on) you face a data migration event.

This has traditionally been do-it-yourself, and has been particularly tricky across Holochain version updates. We've introduced init properties, a new feature that lets you inject arbitrary data at app install time — it's still up to you to decide what that data is and how it's migrated into a new source chain, but this component gives first-class support to some of the necessary plumbing.
This feature is still in preview mode; the dev tooling is still in the works. We're releasing this feature because Unyt needs it now — as a flagship app that's been tested and is rapidly approaching maturity, Unyt needs a mature, straightforward user experience across beta Holochain releases so it can iterate fast while keeping user data intact.

Click Here for the Entire Blog!

Are you a Developer? Join Holochain.org’s Developer Portal: https://developer.holochain.org/

For a complete listing of all of the Holo Ltd and Holochain Foundation events go to their ‘Upcoming Events’ page!

🌟Other Related News & Events

Article: Is private data processing PIE in the sky? Compute without custody by Mary Camacho

The core of Cirdia's Private Intelligence Engine is now published in the Technical Disclosure Commons. How to compute on the data people confide without ever holding it, on ordinary commodity cloud.

Request for Suggestions: All Tech Is Human's Responsible Tech Guide 2026

We're developing our next Responsible Tech Guide and would love your suggestions about people, organizations, and resources to feature! First released it in Fall of 2020, All Tech Is Human's Responsible Tech Guide is read by a wide variety of individuals across civil society, government, industry, academia, and philanthropy.

This year's Responsible Tech Guide will be released on October 1, 2026 and heavily promoted in our newsletter, LinkedIn, Slack community, and during our upcoming Responsible Tech Summit on October 29, 2026 (in-person in NYC + livestreamed to thousands globally).

Send over your ideas! Deadline for submitting suggestions is Wednesday, September 9, 2026 by 11:59 pm ET.

FELLOWSHIP ALERT: Civic Health Project , forwarded by All Tech is Human

Civic Health Project has a $10,000 unrestricted grant award for a 12-month fellowship program "to support early-stage innovators building AI-powered tools to bridge divides, improve civil discourse, and advance social and civic cohesion." US-based.

Submit your concept by August 14th here.

Event: Monetary Diversity Gathering II in Brussels, October 23-25, 2026

Municipal tokens, bioregional finance, public payment rails. Thousands of these projects exist, but builders and policymakers rarely meet in the same room.
Monetary Diversity Gathering II is trying to fix that: 100+ professionals, including municipal reps, infrastructure providers, researchers, monetary innovators; three days in Brussels, Oct 23–25, 2026, working the question: how can new monetary infrastructures strengthen territorial resilience and public innovation?

Pre-register Here

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

In September 2024, we wrote about the Volla Phone Quintus, a de-Googled smartphone from a small tech company in Remscheid, Germany. The Holochain angle was mostly a promise then: a messenger and a backup app were in development, while Holochain on mobile itself was still finding its feet.

Last month Volla held its eighth Community Days under the motto “Stages on the Journey: from promise to reality, and beyond”. Across two days of talks and workshops, the company showed those Holochain apps (hApps) working on Volla and other Android phones.

• Volla Messages, a private communication Holochain app, reached its v0.12.8 release earlier this month. Audio and video calling will be shipped as a public beta later this July. It is available to download today for Android, Linux, Windows and macOS, with iOS and Ubuntu Touch planned for later in the summer.

• Volla Recovery, a hApp which backs up your phone data to the devices of people you trust, entered public beta in July. Recovery is Android-only, for now.

VollaOS 16 ships with something quietly remarkable: a Holochain conductor, the engine that runs Holochain applications, is installed as a system service on every Volla phone.

What makes that worthy of a closer look is how Volla answered the questions that have dogged peer-to-peer software on phones for a decade: what does constant peer networking do to a battery, how do messages arrive when the app is closed, and who keeps everything in sync? The answer to all three turns out to be the same piece of software, a single Holochain conductor that manages all the hApps on your smartphone.

One Conductor to Run Them All

The Volla Cloud service is activated using a deceptively simple toggle in the OS settings that enables mobile hApps to work.

Ordinarily, every Holochain app runs its own conductor. Each app boots its own engine, holds its own network connections, and does its own synchronising. On a desktop machine that’s acceptable. On a phone, where every background process that’s running compounds battery drain, Android and iOS aggressively close everything they deem idle.

Volla’s answer is to run just one conductor for the whole phone, embedded in VollaOS as a system service that stays alive in the background. Dr. Wurzer pointed to it as the breakthrough,

“This is the essential point, because Holochain was not made for mobile devices. Mobile operating systems strictly shut down processes that are not used by foreground apps.”

Apps don’t need their own engine; they bring only their rules and data model, and install it into the shared system conductor. As it was put at the developer workshop, the app “knocks on the service door and asks it to run the hApp”.

The “cloud” in Volla Cloud is nuanced here; there’s no data centre. Volla describes it as a “cloud without computer centre, built by connected devices”. Each app, and chat or group within it, gets its own distributed hash table (DHT), a shared data store spread across the devices of the people using that app, and nothing else. Your messages live in a DHT shared with your contacts. Your backups live in a separate DHT shared with the friends who host them. Your identity, and the conductor that manages it all, remain on your device.

Volla, true to their user autonomy ethos, has built explicit user opt-ins. When an app first asks to use the Volla Cloud service, a permission prompt appears, the same way Android asks about accessing your camera or microphone. Approved apps show up in settings on a registered list with their own individual on/off switches. Consent is visible and revocable, at the level of the operating system.

Peer Persistence

Peer networking is real work: holding connections open, validating entries, gossiping data to other devices. Multiply that by every hApp on a phone and the arithmetic turns ugly. “It wouldn’t be very efficient to sync four or five different conductors, one for each application,” as Davide Garberi, a Volla core developer, put it during the workshop; the shared service means the work is done once, however many apps use it.

Sync and notifications are two faces of the same problem: what happens when the app is closed? On most de-Googled p2p phones, the honest answer has been “not a lot” due to Android’s background-service attitude, and push notifications that are routed through Google’s servers need re-routing. In eliminating that dependency, Volla built a service that is structurally distributed, even from themselves.

Volla Messages comes in two types of builds: a Lite version for VollaOS, which uses the shared conductor, and a full version for other OSs, which by necessity carry their own conductor. Optimised for their own architecture, Lite starts faster, and the hApp is smaller, with the added benefit that battery use is lower with the more efficient conductor configuration.

That single conductor running as a system service keeps peer connections alive when no app is open, receives messages as they arrive, and raises notifications. They are carried over encrypted connections between devices, and arrive on your phone by the time the notification is tapped. In another nod to user agency, settings are planned that will allow users to configure their own service syncs, scheduling heavy transfers for wi-fi rather than cycling through mobile data.

Peer with Friends

Volla Messages has a novel flow when adding a new contact, and the difference is the point. You share your contact code, in person via QR code or over a channel you already trust, such as Signal; your friend creates the contact and sends back their confirmation code, which you accept. Under the hood, the two devices have exchanged cryptographic keys, giving certainty about who’s at the other end. There’s no user lookup to search, no centralised phonebook to query, because there’s no centralised directory: contacts exist between people because they agreed to know each other.

Each conversation becomes, quite literally, its own DHT network. When you start a chat, your device creates a private space that only invited users can join; the invitation is the key to the room. Volla pitches the private messages as a Signal alternative, and for the Telegram group chat crowd there are open channels, joined with a public invitation, where the conversation itself is carried on the devices of everyone in it. That has consequences for resilience: channels don’t depend on any particular peer or device; instead, members receive chat history from the network of peers that are still in the channel. The team shared, larger groups sync more smoothly, since there is a greater number of distributed devices to fetch from. Asked during the Volla Days presentation about channel size limits, Amirhossein Esmaeilipour, who builds the app’s Tauri side, said he knew of none! Moderation tools, designed to support those open p2p channels, are in development.

Then there’s the question every Holochain-curious reader will ask: if the underlying ledgers are append-only, how do messages get deleted? Volla’s answer is data migration. The team is building migration tooling to survive Holochain version upgrades, which can break DNAs, forcing a new DHT network: the same machinery does double duty here. “We found a solution to remove data,” Dr. Jörg Wurzer, Volla’s founder and CEO, put it in the keynote. When data moves to a new structure, anything marked as removed simply isn’t carried forward into the new network. Speaking to us Dr. Wurzer added, “The solution is versioning the DHTs”, which, like many good ideas, sounds like common sense once realised.

Dr. Wurzer has already found another solution for the app to make life easier: personal infrastructure. He runs a private one-person network between his Volla phone and his MacBook, using Messages to sync his clipboard and shuttle files, “more easy than connecting with a cable“. Copying and pasting between devices got a whole lot easier despite their divergent operating systems. A messenger with no server turns out to be a rather good pipe.

Volla Recovery answers the question every de-Googled phone owner eventually faces: where do backups go when Google’s cloud isn’t an option? Volla’s answer: to the devices of people you trust. Settings, contacts, calendars, messages, apps, photos, music, documents; you tick the sources, invite a friend with a code, and their device becomes your backup host. Hosts stay in control, allocating how much storage to lend with a simple slider. Any Android phone can host the backups, so that old handset in a drawer, plugged in behind the router, could become the family storage node. Crucially, those host devices store everything and can read none of it, meaning backups stay private.

Security responsibility is stated plainly. Your recovery keys are derived from a password, and there is no reset. “If you lose your password, your data is gone, and that’s it,” Garberi told the room. “There is no recovery at all, because otherwise it wouldn’t be as secure.”

Restoring data from hosts demonstrates the peer-to-peer topology. There’s no restore wizard with a progress bar; your data returns as your chosen hosts come online, gossiped back in piece-by-piece. Which points clearly to the clever part. Rather than every host holding a full copy of all your files, backups will be split into chunks, each stored on two or three hosts. Automatic health checks will add a latent resilience: if a host disappears for a few days, its chunks are quietly re-replicated to another host. Dr. Wurzer notes this functionality runs ahead of the Holochain framework itself: “the Holochain Foundation has sharding on its roadmap, but it’s not implemented yet, so we had to find our own solution”. Open source working as intended; the code sits in Volla’s public repositories with Holochain Foundation core and community developers committing alongside. Both benefit from the stimulus of necessary invention.

The deeper promise is what’s absent. “There’s no Volla account at all,” Dr. Wurzer said, and there won’t be one: “the last that we know about our devices is that DHL has sent it.” Backups exist in a web of people that you chose and nowhere else.

Peer Together

The conference’s promises came with receipts. “Everyone is invited from the community to review the code, to contribute.” The Volla Cloud app went public the day of the workshop, and the modular open-source CAD files for their latest phone hardware the day after, with Recovery‘s full source following in July.

For hApp builders, there are two routes in:

1. Web-stack developers can follow the Messages pattern, SvelteKit inside a Tauri shell, and simply swap one plugin to use the shared conductor instead of bundling their own.

2. Native Android developers can follow Recovery’s path, calling the service from Kotlin.

Either way, the finished app asks the user’s permission, appears in the Volla Cloud Service’s registered list, and inherits the sync, notifications and battery behaviour described above.

What Volla is offering the Holochain ecosystem is a scarce thing: distribution. Phones ship with a conductor already running; a nightly app store channel exists in Volla’s F-Droid repository. A Google Play listing has been submitted and is currently under review, while an iOS App Store listing is planned for later this summer.

The invitation extends to hardware. Volla publishes their hardware designs, so anyone can design accessories that attach to the magnetic mounts on their latest phone, the Plinius. Volla’s own first module, the Cell Mini radio, turns it into what the company calls “the world’s first meshnet phone” and is crowdfunding this autumn.

Dr. Wurzer spelled out what that capability promises, “People can run hApps or Volla Cloud apps even in case of blackout or internet restrictions… an infrastructure for people by people, free to use. This will give people full sovereignty.”

An ecosystem that opens its device schematics as readily as its source code is making an argument about trust, both in the robustness of their design and in the community they are inviting.

Understated Competence

What shipped in June is deliberately modest: a toggle in settings, a messenger, a backup app. What it demonstrates is not modest at all: Holochain running as consumer infrastructure, on ordinary phones, for people who don’t care what a conductor is.

Those hApps are available today. Messages is a free download for the phone or computer you already own; no account, remember, just a contact code shared with a friend. The fuller experience, the always-on conductor, Recovery, the battery-friendly Lite apps, comes with the phones themselves available at volla.online.

Useful Links

Volla
Volla GitHub
Volla Community Days 2026, Day 1 stream
Volla Community Days 2026, Day 2 stream
Keynote slides (PDF)
Volla Cloud apps workshop slides (PDF)
Volla Cloud Services
Volla Message
Volla Recovery

🌟 Latest News from hAppenings.community

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

🌟 Latest Official News and Events from Holochain, Holo & Unyt

Event: Strategic Update Livestream on Thursday, July 30th at 14:00 UTC, from the Holochain Foundation, Unyt and Holo Hosting

On July 30th, we are hosting our next Strategic Update Livestream. Following up on the massive technical milestones and coordination we saw in the first half of the year across Holochain, Holo Hosting, and Unyt Accounting, the team will be taking some time to discuss our ongoing deliveries, current network momentum, and what’s next on the horizon.

If you are a developer, an ecosystem partner, or just someone passionate about peer-to-peer technology, we’d love for you to join the conversation.

Set a notification to join us live on YouTube!

Are you a Developer? Join Holochain.org’s Developer Portal: https://developer.holochain.org/

For a complete listing of all of the Holo Ltd and Holochain Foundation events go to their ‘Upcoming Events’ page!

🌟Latest News from the Holochain Ecosystem:

Blog: Introducing a New Era of Community-Directed Liquidity from Neighbourhoods

At Neighbourhoods, we’ve always believed in creating infrastructure that empowers communities — giving them the tools to coordinate, govern, and thrive on their own terms. Today, we’re excited to announce a major step forward in this mission with the introduction of Neighbourhoods Liquidity Nodes (NLNs): a bold shift towards community-directed liquidity, powered by decentralised protocols that are aligned with neighbourhoods infrastructure, not legacy intermediaries.

Click Here for the Entire Blog

🌟Other Related News & Events

Announcement: Tapestry Application now available from Cirdia, Mary Camacho-Founder

Every day gives you another piece of the picture.

The breakthrough isn’t in a single entry - it’s in the connections that emerge over time.

That’s why Tapestry delivers clear weekly insights, helping you understand your wellbeing with confidence.

Start your free trial today!

Event: Wikimania 2026 - Paris, France July 21-26, 2026 from the Internet Archive Europe.

Looking forward to Wikimania Paris this week and celebrating 25 years of Wikipedia and the incredible community that has built up this invaluable resource.

Together with Daniel Erasmus, Internet Archive Europe will be showcasing a new tool still in development to help strengthen Wikipedia citations via analysis of academic papers. If you’re at the conference, please come by on Wednesday morning at 9:30, otherwise see us around, our Program Manager Beatrice Murch will be here all week.

Event: DWeb Camp Cascadia - Salt Spring Island, Canada July 30-August 3, 2026

The decentralized web movement is global, and our summer schedule reflects that! Right on the heels of DWebCamp Berlin, one of our team will be heading to DWeb Cascadia from July 30th to August 3rd.

Building a truly peer-to-peer internet requires gathering, sharing ideas, and collaborating in person. If you’re going to be in the Pacific Northwest for the camp, we’d love to connect and discuss agent-centric architecture, p2p networking, and what you’re building.

Let us know if you’ll be there! Tickets are still available! Find out more about the camp here: https://dwebyvr.org/camp/

Event: Monetary Diversity Gathering II in Brussels, October 23-25, 2026

Municipal tokens, bioregional finance, public payment rails. Thousands of these projects exist, but builders and policymakers rarely meet in the same room.
Monetary Diversity Gathering II is trying to fix that: 100+ professionals, including municipal reps, infrastructure providers, researchers, monetary innovators; three days in Brussels, Oct 23–25, 2026, working the question: how can new monetary infrastructures strengthen territorial resilience and public innovation?

Pre-register Here

Article: Lock the Web Open - Inside DWeb Camp 2026 from the Internet Archive Europe

“Lock the web open.”

That was brewster kahle‘s closing line as DWeb Camp 2026 gathered more than 500 builders, researchers, archivists, and artists in a forest outside Berlin to imagine what a web built for the public interest could look like.

The phrase stuck because it challenges a common assumption. An open web isn’t something that maintains itself. It has to be designed that way, supported by infrastructure that isn’t controlled by a single platform, and protected before someone else decides who gets access.

Across five days of conversations, workshops, and late-night discussions, the same idea kept resurfacing. Whether the topic was public AI, community-run infrastructure, local-first technologies, or digital preservation, the goal wasn’t decentralisation for its own sake. It was resilience: ensuring that knowledge, culture, and public memory remain accessible regardless of who owns the next platform.

For Internet Archive Europe, that’s why DWeb Community matters. A decentralised web and a legally protected right to digital memory are different approaches to the same question: how do we keep knowledge open for everyone?
***
Photo credit: Anton Tal

Film - The Wicked Mush of the In-Between with Bonnitta Roy, documentary by Katie Teague

About 3 years ago I (Katie Teague) started my "In the Making" series as a side project. It started with my interview with Bonnitta Roy, founder of Alderlore Insight Center. This short (10 mins) film is from that interview. It recently resurfaced and I rewatched it and realized it is one of my favorite of all that I've crafted these last few years.

In this conversation, Bonnie asks what if the "metacrisis" isn't a crisis at all, but a once-in-a-civilization transition (a time between worlds)? She invites us to see our moment not as something to fix or escape, but as the "wicked mush" of the chrysalis, where old systems dissolve and new possibilities emerge. Drawing on the Panarchy framework, we explore uncertainty, imagination, and why the most important work of our time may be learning to inhabit the in-between...not as a place to fear, but as the fertile ground where new worlds begin.

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

Flowsta’s Vault is a free, open-source desktop app that puts your digital identity on your own machine. We’ve followed Flowsta from its login button to ProofPoll via building our own app Fieldnotes. This time, we opened the code to see if Vault does what it says, and watched its v1.0.1 release land as we went to press.

Where Your Identity Lives

Most of us have spent years choosing passwords built from a pet’s name, a film character, or the same word with a number on the end. Password managers improved matters by generating strong, randomised codes for you. Flowsta leans on the same idea that made password managers work: a small app, on your own device, with a clean interface and sensible guardrails, doing the hard part for you. It is a deliberate answer to the objection that has held decentralised computing back. People are willing to control their own identity; they are not willing to become cryptographers to do it. By making self-custody feel ordinary, Flowsta removes the fear without removing the control.

Your 24-word seed phrase

When you sign in to most online services, your password is checked against a copy held on the company’s cloud server. Their server is the authority. It decides whether your credentials are acceptable and whether to grant you access to your data. If their server goes down, access to your own data goes with it.

Vault, from Flowsta, works the other way around. During setup, Vault generates a unique 24-word seed phrase for you, on your own device, which acts as your identity ‘seed’ and recovery phrase. Your device uses it to generate a key pair, split between a public key you share and a private key you keep secret. That phrase alone can restore your identity, so it should always be kept separate, secure and private. As of yesterday’s v1.0.1 release, a brand-new identity can even be created with no internet connection at all; Vault simply registers it with the network the next time you’re online.

That ability to reliably generate and control your own keypair shifts the centre of gravity to you as the owner and keeper of your identity. Since there’s nothing on Flowsta’s servers to leak or lose, they can’t help with password resets, and equally can’t withhold your data, whether through an accidental outage or a deliberate restriction.

Vault also derives a decentralised identifier, or DID, from your public key, following the W3C‘s open standard. Flowsta doesn’t issue it and can’t revoke it; it’s yours because the keys are. It can be used across the growing decentralised web.

All Your Keys

If your unique identity is being kept safe by Vault on your laptop, for example, how do you retain coherence using the same Holochain application (hApp) or website on another device? Flowsta’s answer is to use the 24-word phrase to generate multiple key pairs: one for each device you link to your Vault. To prove they belong to the same person, a short statement is signed by both sets of keys, which only you could have produced with your phrase. The link stands as a mathematical truth.

Sign It as Yours

Once your keys live on your device, they can be used for more than signing into an app. They can indelibly mark any file as yours. Vault’s ‘Sign It’ feature lets you sign a file with your private key and, alongside the signature, attach a plain-language statement detailing what the file is and how it can be used. The statement is machine-readable, so other software can act on your choices, like the licence, and whether AI can use it for training.

The file itself never leaves your device. Vault reads it and produces a mathematical fingerprint that changes completely if even a single byte of the file changes. Your key signs the fingerprint, and only the fingerprint, your signature, and your statement are written to a shared public record that uses Holochain’s distributed hash table (DHT) as a ledger held across the users’ own devices rather than on a Flowsta server.

A small preview thumbnail of a few kilobytes accompanies the record, so it’s recognisable at a glance. It’s generated automatically as a shrunken image, an audio waveform, or a document’s first page, and you can swap it for a custom one you feel suits it better.

Images get a second kind of fingerprint. Exact fingerprints are unforgiving of images: re-save a photo in a slightly different format and it no longer matches even though the content is the same. So for images, Sign It’ also records what the picture looks like, a fingerprint that survives resizing and re-saving, so a lightly altered copy can still be traced back to the original.

Anyone can check a fingerprinted file against Flowsta’s Verify website to see if there’s a matching record. Because your signature can be tested against your public key, the validation rests on maths, not Flowsta. Exact copies return a verified mark; resized or re-edited images come back with a percentage score showing how similar they are.

You’re not bound to a signature forever either, despite Holochain’s immutable architecture. You can replace a signature with a new one under different terms, which points back to the one it has superseded, or mark it as revoked, giving an optional reason. The account of what you have signed, varied, or withdrawn stays honest, because every step is public and permanent.

Strong Room Ledger

The strongest thing about Vault is not any single privacy or security claim. It’s that you can check the claims for yourself by looking at the published, open-source code. We did that and can say: your keys, your recovery phrase, and your private data never leave your device unencrypted.

What is shared? While Vault is unlocked, it keeps a simple monthly tally: for each app you use, one record a month with a first-seen date, a last-seen date, and a use counter. That’s Flowsta’s MAU, the monthly-active-user count their billing runs on. Use three apps and that’s three records a month, no more. It’s less a usage feed than a monthly attendance sheet that syncs when it gets the chance; Vault checks every half hour whether there’s anything new to send, and usually there isn’t. Each record carries a random per-app code, generated once on your device, and nothing else: no passwords, private keys, files, email, or username.

Two things are worth noting. Because each app’s code is random and its own, one developer’s numbers can never be lined up against another’s. The only account-level identifier is separate: a one-way hash whose only job is to point Flowsta at the right identity for billing, and it can’t be turned back into your keys or phrase. That identity is only a cryptographic one, keys and a hash, with everything personal behind it, your data, your files, your email, encrypted in your Vault, on your device. Flowsta can count that it was active, to operate their plans, which include free tiers as well as Premium plans that are priced per active user, offering a higher monthly ‘Sign It’ allowance and shorter public identity names. What they can’t do is look inside it.

To a developer, all of this arrives as a single aggregate number: we built Fieldnotes on top of ProofPoll, and we can see how many people use it, but nothing about the users' other apps or Flowsta services. That's the zero-knowledge design working: Flowsta can count that an account was active, but they can't look in, and the cryptographic identity only becomes "you" if you choose to make the link, by claiming a public username or sharing your DID. The records travel to Flowsta over HTTPS and go nowhere else, no public network, no DHT, no third party. It's the smallest possible data surface that still supports a reliable service, and Flowsta have published a field-by-field account of exactly what Vault sends.

Stronger Together

To join the ‘Sign It’ network, ‘Vault’ uses Flowsta’s bootstrap servers for discovery of others in the peer-to-peer network; and as of v1.0.1 it can also fall back to open, community-run bootstrap servers if Flowsta’s are unreachable. The same is true when a file is checked on the ‘Verify’ website: the maths does the verifying, but the page that runs it, and the identity records behind each signature, are Flowsta’s. The revocation mentioned earlier lives here too: whether a withdrawn signature is shown or hidden is an interface choice, not something the network enforces.

Eric set himself a test: what if Flowsta disappeared tomorrow, could users continue to access their own data? What’s not in doubt, what’s already on your device stays: your keys, encrypted data, and your identity are only ever on your device, so there’s nothing there to seize, leak, or lose. Your installed Vault keeps opening, and your local data remains readable. What would currently falter is Flowsta’s added services beyond your device. Naturally, updates to Flowsta’s own software, Vault, and the three Holochain applications (DNAs) beneath it would stop, and the hosted Verify page would go dark too. Peers could, if the necessity arose, spin up a verification explorer website to continue to surface the ‘Sign It’ network data. So the promise that holds: your identity is yours, and it survives. The second half, a network that runs whether Flowsta is there or not, is still a direction of travel toward peer resilience.

Your Vault

We drafted this article about the road ahead, and while we were editing, Eric shipped a stretch of it. Yesterday’s v1.0.1 release makes the Flowsta Vault desktop app the front door to your sovereign identity. New identities are now created in Vault itself, and signing in with Flowsta on the web asks your installed Vault for approval rather than a password.

Existing web accounts are walked through an easy migration; until now, Flowsta’s servers held the keys for web-only accounts, an early custodial setup that growing Vault adoption had outgrown.

The same release ships the storage layer for syncing between your own devices: an encrypted, per-user record the network passes around without being able to read it.

In Eric’s words, “Vault isn’t a client phoning home to us, it’s a real node in the network... the more identity lives on devices, the less the whole thing depends on us at all.”

Sharing Flowsta’s own node work, in a secure manner, across a distributed network of nodes is the goal, and 1.0 puts the first big piece of it in users’ hands.

Further out sits the reason much of this is being built now. Flowsta is developing something it calls ‘Your Own AI’, and Vault is the identity layer beneath it. Authentication came first because the team needed it for the AI platform. What that becomes isn’t yet ready to share, but it is the next thing we’ll be asking Eric about!

Verify Vault

Most services ask you to blindly trust them. Vault proves the opposite, that you shouldn’t have to; your keys should sit on your own machine, and the codebase should be open. For every claim we’ve made here, you can check against that code yourself. That’s the real offer. Not a promise to handle your data well, but a design where the promise isn’t needed.

We’ve checked it because the software invites it. Much of what Flowsta claims holds true without need for qualification: your keys and your data are yours, encrypted on your device, and no one at Flowsta can read them or take them away. Usage counts do reach Flowsta’s servers, and discovery, ‘Sign It’ verification, and app updates still route through their servers today. You have sovereignty over what you already hold. A network that depends on no one, Flowsta included, is still being built.

This is what building in the open looks like, and it is the best reason to take Flowsta seriously. A company that publishes the code revealing its own dependencies is not asking to be believed. It is asking to be checked and standing by its methods.

Vault is free and open source for macOS, Windows, and Linux, with mobile on the way. Its codebase is public for you to inspect before you install it. If you’re a developer building a decentralised application, Flowsta’s platform offers a comprehensive solution that’s easy to deploy and integrate. Their Discord server is a useful place to connect with the team and work through any questions you might have as a user or developer.

If you want your identity under your control, on your machine, rather than on someone else’s server Flowsta’s Vault is the home for it.

Download Flowsta Vault

Useful Links
Flowsta
Flowsta:Dev
GitHub
LinkedIn
X
Facebook
YouTube
Discord

🌟 Latest News from hAppenings.community

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

🌟 Latest Official News and Events from Holochain, Holo & Unyt

Blog: You Can’t Run an Economy on a Bet by Cameron Burgess from Holo.host

(This article) explores why the decentralized web has to move past speculative “casino” mechanics to build stable, predictable economies. For the devs in here, this is a great reminder of what kind of currency design is right for running your ecosystem’s marketplace — it shouldn’t bounce like a yo-yo.

Instead, it should be anchored to actual productive capacity.

That’s what Holo and Unyt have built together for HoloFuel.

Note: Here’s a quote from the article that’s of particular importance:

We’ll name the date of the real HOT-HF Technical Migration on a livestream, July 30th.

Join our Holo Telegram community to stay updated and join the wider the conversations. In order to support those conversations we just updated our Telegram channel to multiple topics including: Holo Hosting, Unyt, Holochain, Moss, and General. Looking forward to seeing you in a topic there soon! https://t.me/channelholo

Click Here for the Entire Blog

Holo Roadmap Update: Q2 milestones achieved, setting the stage for Q3.

We are excited to share an update on our product roadmap. As you can see, Q2 was a period of massive foundational delivery. We successfully completed:

✅ The HOT to HoloFuel Technical Migration Test

✅ The Holo Web Conductor Beta

✅ The H2HC Linker Beta

As we move into Q3, our focus shifts toward direct utility and economics for our network hosts. Upcoming deliverables include the HOT Hosting Utility Integration and enabling Hosts to earn HoloFuel for Wind Tunnel and Always-on Node Hosting.

We also have a major new milestone announcement coming soon. Thank you to our community and hosts for building alongside us!

Review the full roadmap here: https://holo.host/product/roadmap/

Announcement: We are heading to Germany for DWeb Camp 2026! - Holochain Foundation

The decentralized web is built on strong communities and shared infrastructure. From sponsoring the event in the past to attending year after year, DWebCamp has always been a special gathering for the Holochain team.

This year, the camp is moving to Europe, and we are thrilled to announce we will be attending DWebCamp in Berlin, Germany!

From July 8th to 12th, we will be on the ground connecting with fellow builders, researchers, and decentralized tech advocates. If you are attending and want to talk about agent-centric architecture, peer-to-peer applications, or just want to say hello, we would love to connect with you.

Learn more here!

Are you a Developer? Join Holochain.org’s Developer Portal: https://developer.holochain.org/

For a complete listing of all of the Holo Ltd and Holochain Foundation events go to their ‘Upcoming Events’ page!

🌟Latest News from the Holochain Ecosystem:

Blog: Solstice Update (Well, Kinda Still Solstice) from the team at HummHive

Welcome to the Winter Solstice Update from Australia and of Course the Summer Solstice to Our Northern Hemisphere Friends. A great time to be updating you on our progress and story. Here is what's been going on with HummHive over the last month.

Click Here for More

Update: Volla Messages by Jorg Wurzer from Volla

The planned EU chat regulations make it clear: developing Volla Messages was the right decision. Volla Messages uses the decentralised Volla Cloud powered by Holochain, which has no data centre. Every communication is solely a matter for the people who choose to get in touch with one another, both in terms of content and technology.

Volla Messages is currently available as a public beta for Android, Linux, Windows and macOS. The stable version is scheduled for release later this month. Under Volla OS, Volla Messages uses a system service that maintains the connection in an efficient manner.

However, privacy and security are only fully guaranteed when Volla Messages is run on a secure device. If the Gboard from stock Android is reading along, communication is being monitored. That is why we offer Volla Messages on our Volla Phones and tablets, running Google-free Android and without any backdoors.

For more on the Volla Cloud, click here.

🌟Other Related News & Events

Resource: Database: Financing Alternative Ownership Enterprises from Transform Finance

Welcome to the Transform Finance database of financing vehicles for Alternative Ownership Enterprises (AOEs) — firms that significantly shift economic value and decision-making power toward the non-investor stakeholders they impact, such as workers, producers, consumers, community members, or even a non-financial purpose. Employee ownership is the most active area of financing, but the database also includes financing for all cooperatives, steward ownership, and other models.

This database aims to provide a starting point for investors seeking funds to back, and for advisors and businesses seeking capital for transactions. The most recent update is July 2026 with an analysis of new funds and trends here.

Blog: UN-ADVERTISING - What happens when AI unbundles advertising, marketing & PR into data and code — and re-bundles them around proof? by Scott Frankum

The semantic web uses structured data — shared vocabularies foundational to AI and blockchains — that transforms simple search into novel automated services. We’re in a new reality, where earning the buy box is a formula: Context Completeness × Trust.

I argue that agentic commerce reduces everything except desire, taste, identity, belonging, status, meaning, and fandom—to math (expressed as data and code). You used to build trust. Now you’ll sell receipts (attestation and proof).

Click Here for the Full Article

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

(This is a shorter version - if you’d like to read the full length article read it here)

“Build your own app” has always carried a hidden instruction: learn to code first. It was never optional. You could have ideas about software, but making it was someone else’s job, behind a wall that took years and money to climb. Ideas on one side, the ability to build on the other. Flowsta recently released ProofPoll, and their blog promises “it’s the fastest way to build a Holochain app”.

I’m not a programmer. I don’t write code in any language; Anita and I made this newsletter understandable for people who aren’t developers. This week, I built a working Holochain application (hApp): not a mock-up, not a demo that collapses the moment you lean on it, but a real peer-to-peer app with its rules enforced at the core. I didn’t hand-write the code, and I didn’t set up the automated tooling a developer might use. What I had was a sound starting point, ProofPoll, and an AI assistant.

The wall is lower. It hasn’t vanished, and that difference is the whole point: it’s where we choose to build humane, agent-centric software.

Proof in the Field

The new app is called Fieldnotes. It runs structured test scenarios: someone sets out a list of things to try, testers work through them, recording a verdict on each and adding observations and attachments as they go. It grew out of a real need of our own, coordinating community testing for the Requests & Offers hApp. Fieldnotes isn’t the point; it is the evidence, and the code is public on GitHub.

I started Tuesday morning with an idea; by the small hours of Wednesday, there was a working app, and later that day I released an installable version under an open licence. In between was a full night’s sleep, the school run, and a supermarket shop. Flowsta says on their blog that they’ve watched a working app come together in an afternoon, with ProofPoll and an AI coding tool. From idea to released hApp took me just over a day.

I want to be clear, because that is the kind of claim that invites ridicule. It wasn’t a day because building peer-to-peer software has become trivial. It was a day because almost all of the software was already built, soundly, by other people, and what was left for me was small.

ProofPoll

The complete application I started with is ProofPoll, built by Flowsta. It runs polls and votes that can’t be faked or quietly altered, and it’s free for anyone to copy, take apart and rebuild. That last part is deliberate: ProofPoll is built as a template, the hard problems solved once, in the open so that the next person can point them at a different purpose.

To see how much was already done, look at the project’s history. Software built with Git keeps a dated record of every change ever made to it, and because Fieldnotes grew out of ProofPoll, it carries that whole history with it. It opens on the first of March 2026: identity, encryption, the update system, release configurations, three months of work, all by other people. My addition is the shortest stretch at the very end.

I didn’t touch the identity system: signing in, proving you’re the same person across two devices, keeping your keys safe. The fiddly business of running a Holochain application on a small server inside your own desktop, and keeping it alive through crashes, comes from Flowsta’s Vault. So does private note-keeping; so do the pipelines that publish the Mac app. What was left for me is small enough to say in a sentence: I changed what the app is about, from polls and votes to test scenarios and findings. I built a handful of things on top, because the core complexity was already taken care of.

One piece didn’t transfer: ProofPoll’s published versions carry Flowsta’s Apple-issued code-signing certificate. I inherited the app code, but not the certificate, so Fieldnotes ships unsigned. Mac is the only build released so far; if you run it, your Mac will warn you the first time you open it. Right-click the app in Finder, choose ‘Open’, and after that, it should behave normally. The full step-by-step is in the repository.

Proof Positive

I’ve been honest about what I inherited, and I’ll talk about the problems. First, I want to be precise about the part that gave me no friction at all.

Flowsta has built a full suite of tools for handling cryptographic identities on the decentralised web. When Fieldnotes needed an identity system, it leaned entirely on Flowsta’s, and so did I, as a user. I created an account for free, linked it to the app, and re-authorised through Flowsta’s login each time I rebuilt. It was, in the most ordinary and welcome sense, faultless.

That matters more than it sounds. Identity is foundational to distributed systems like Holochain, and one of the hardest things to make pleasant. Most peer-to-peer systems ask you to understand something technical before you can begin. Flowsta asked me to use things I already know: single sign-on, two-factor codes, and a password manager. The difficult cryptographic work doesn’t disappear; it moves somewhere I don’t need to think about, as either a user or a developer. Doing that quietly is what makes Flowsta’s work special.

Absolute Proof

Here is the thing nobody tells you about building this way. The AI is astonishingly good and, intermittently, confidently wrong, and the two are not always separable. It is the brilliant colleague who has read every book in the library and never once set foot in the lab. Ask it the general shape of a thing, and it’s usually right. Ask it the exact name of an unusual function, or the precise behaviour of a tool it has never actually run, and it will make things up with total conviction. More than once, it sent me down a path that didn’t exist, described as confidently as the paths that did.

You can’t always tell which kind of answer you’re getting. So the whole discipline comes down to one habit: never trust the claim, check the source. AI is wonderful at proposing solutions and unreliable at remembering exact context; let it reason from memory, and it will eventually, charmingly, walk you off a cliff. What checking looks like in practice is in the companion guide.

The clearest example is also the most important part of the whole app. Fieldnotes has a lock on it: only an authorised administrator should be able to add test scenarios or view screenshots, and that rule is meant to be enforced at the deepest level, not just hidden behind buttons. The AI and I built it as a variation of ProofPoll’s lock. It looked right; the app complied and the UI behaved exactly as I expected. It wasn’t actually locking. It was letting things through while silently failing. In an app demonstrating secure identity, that’s a problem.

What caught it was not a flash of insight. It was a test. I needed to check two people on the same network, one authorised and one not. I’m one person with one computer, so I can’t be two genuinely different peers without a test harness. I couldn’t have written it myself, so I asked for it. Fieldnotes failed the test: the unauthorised person was getting in. I traced it down, fixed it, and ran the test again. This time, the unauthorised peer was refused, and even an attempt to slip through by faking someone else’s permission was prevented. The lock held, and now there’s proof it held. That check is what stands between “it looks like it works” and “it works”; between those two is exactly where confident, untested answers live.

The tool that caught it is Holochain’s sweettest, runs checks across several simulated peers. The lesson is simple: run tests as you build, and before you share. You don’t have to write them by hand, but you do have to know to ask. The companion shows how, without a word of code.

Mutual Proof

“A non-developer built a hApp in a day” hides more than it tells. Others did the heavy lifting, and there were a few things I leaned on.

• Holochain Agent Skill, written by Sacha Pignot, lead developer at hAppenings Community, taught the AI how Holochain actually works before we began: its architecture, its validation rules, and how to test it across peers. It reasoned like something fluent in this world rather than guessing from general web habits, and when it later knew to reach for the right kind of test, that was Sacha’s groundwork showing.

• Claude Opus 4.8 brought speed, recall, and the ability to read code faster than any person, alongside its real and constant liability.

And I brought two things only a person could: judgment about where a peer-to-peer system stops behaving like the software my instincts expect, and the discipline to make the AI prove every claim against the source.

Every item on that list is freely available, and that’s the point. I took the time to think in a particular way, starting from privacy and resilience rather than bolting them on, treating people’s data as theirs to keep. That’s the whole of it, and it’s a smaller thing to ask of someone than “become a programmer first.”

🌟 Latest News from hAppenings.community

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

🌟 Latest Official News and Events from Holochain, Holo & Unyt

Blog: DevPulse 154, by Paul D’Aoust

"Smooth like butter." That's the verdict on the newly released Holochain 0.6.1. 🧈

We've just published Dev Pulse 154, detailing the massive engineering effort behind this release. By integrating the Iroh networking layer into our active 0.6 branch, we've delivered a new baseline for peer-to-peer reliability. Connection drops are minimized, peer discovery is faster, and the network recovers gracefully from chaos.

If you are building agent-centric distributed apps, this is the architecture you want to be building on. Dive into the technical details (and the breaking changes from 0.6.0) in the latest Dev Pulse.

Ecosystem News: Iroh 1.0 Announcements

Huge congratulations to the Iroh team! They just officially released Iroh 1.0. For the builders here, you already know that Iroh is the new networking engine powering Holochain 0.6.1. Their approach to QUIC multipath, hole-punching, and local-first discovery is a huge reason why the recent HOT-HF technical migration test ran so smoothly. If you are interested in the lower-level networking tech that makes Holochain tick, we highly recommend reading their 1.0 announcement blog: https://www.iroh.computer/blog/v1 We also encourage you to drop into their community at their Iroh Discord: https://discord.com/invite/DpmJgtU7cW

Are you a Developer? Join Holochain.org’s Developer Portal: https://developer.holochain.org/

For a complete listing of all of the Holo Ltd and Holochain Foundation events go to their ‘Upcoming Events’ page!

🌟Latest News from the Holochain Ecosystem:

Conference Update: Superintelligence for Humanity, by Matthew Schutte

This research summit on artificial intelligence was full of brilliant insights. I particularly enjoyed some of the comments from Ahmer Inam, Blaise Agüera y Arcas, Pierre-Yves Oudeyer, Divya Chander, Michael Levin, Olaf Witkowski, Damián Blasi and a longtime favorite, Bruce Schneier.

Note: Watch the full inaugural event recording!

🌟Other Related News & Events

Announcement: Emerging Voices in AI & Society Fellows Program from the Center for Humane Technology

Center for Humane Technology is pleased to announce we’re opening applications for our inaugural Emerging Voices in AI & Society Fellows Program.

The cohort will research and help shape public discourse across multiple potential areas, including: AI and Cognition, AI and Relationships, AI and Surveillance, AI and Spirituality, and Inside the Machine: a technologist's view of what drives decisions at leading AI companies.

Applications are open through July 12, 2026. The cohort begins September 14.

If you've been thinking about how AI is reshaping human life and you have something to say that is missing from the current conversation, we'd love to hear from you.

Article: Most of the Renaissance Has Never Been Translated. Source Library Is Opening It. from the Internet Archive EU

Most of the Renaissance has never been translated.

Around 90% of Renaissance Latin has never been translated into a modern language. More Latin was written after 1500 than survives from all of ancient Rome, yet much of it remains accessible only to specialists.

Last week, Internet Archive Europe attended the launch of Source Library at Amsterdam's Embassy of the Free Mind.

With more than 15,000 historical works, 55 languages, and 6,000 first-ever English translations, Source Library is opening a vast body of knowledge that has long been out of reach for most readers.

The project raises a question we encounter often: who gets access to knowledge, and on what terms?

Event: DWeb Camp 2026: Root Systems, July 8-12, 2026 in Alte Hölle, Germany by DWeb

DWeb Camp 2026: Root Systems is a five-day gathering in nature for builders and dreamers dedicated to building a decentralized web (DWeb). A web that is private, resilient, and beyond the control of a few central powers. Like forests, decentralized systems derive strength from what lies beneath the surface: dense networks of roots, sharing resources without hierarchy, and coordination that persists even when individual nodes disappear.

DWeb Camp: Root Systems invites participants to build technologies that–like living ecosystems–adapt, regenerate, and endure.

Note: Here’s another link to DWeb Camp 2026!

Announcement: Dweb Camp Cascadia 2026 call for proposals!

We just announced our theme Co-Adapt, Co-Create, Co-Here. Now we invite you to join us in co-creating DWeb Camp Cascadia, happening from July 30th to August 3rd 2026. If you have an idea for a talk, workshop or any other activity, let us know, we are looking forward to hearing about it!

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

“Build your own app” has always carried a hidden instruction: learn to code first. It was never optional. You could have ideas about software, but making it was someone else’s job, behind a wall that took years and money to climb. Ideas on one side, the ability to build on the other. Flowsta recently released ProofPoll, and their blog promises “it’s the fastest way to build a Holochain app”.

I’m not a programmer. I don’t write code in any language; Anita and I made this newsletter understandable for people who aren’t developers. This week, I built a working Holochain application (hApp): not a mock-up, not a demo that collapses the moment you lean on it, but a real peer-to-peer app with its rules enforced at the core. I didn’t hand-write the code, and I didn’t set up the automated tooling a developer might use. What I had was a sound starting point, ProofPoll, and an AI assistant.

The wall is lower. It hasn’t vanished, and that difference is the whole point: it’s where we choose to build humane, agent-centric software.

Proof in the Field

The new app is called Fieldnotes. It runs structured test scenarios: someone sets out a list of things to try, testers work through them, recording a verdict on each and adding observations and attachments as they go. It grew out of a real need of our own, coordinating community testing for the Requests & Offers hApp. Fieldnotes isn’t the point; it is the evidence, and the code is public on GitHub.

I started Tuesday morning with an idea; by the small hours of Wednesday, there was a working app, and later that day I released an installable version under an open licence. In between was a full night’s sleep, the school run, and a supermarket shop. Flowsta says on their blog that they’ve watched a working app come together in an afternoon, with ProofPoll and an AI coding tool. From idea to released hApp took me just over a day.

I want to be clear, because that is the kind of claim that invites ridicule. It wasn’t a day because building peer-to-peer software has become trivial. It was a day because almost all of the software was already built, soundly, by other people, and what was left for me was small.

ProofPoll

The complete application I started with is ProofPoll, built by Flowsta. It runs polls and votes that can’t be faked or quietly altered, and it’s free for anyone to copy, take apart and rebuild. That last part is deliberate: ProofPoll is built as a template, the hard problems solved once, in the open so that the next person can point them at a different purpose.

To see how much was already done, look at the project’s history. Software built with Git keeps a dated record of every change ever made to it, and because Fieldnotes grew out of ProofPoll, it carries that whole history with it. It opens on the first of March 2026: identity, encryption, the update system, release configurations, three months of work, all by other people. My addition is the shortest stretch at the very end.

I didn’t touch the identity system: signing in, proving you’re the same person across two devices, keeping your keys safe. The fiddly business of running a Holochain application on a small server inside your own desktop, and keeping it alive through crashes, comes from Flowsta’s Vault. So does private note-keeping; so do the pipelines that publish the Mac app. What was left for me is small enough to say in a sentence: I changed what the app is about, from polls and votes to test scenarios and findings. I built a handful of things on top, because the core complexity was already taken care of.

One piece didn’t transfer: ProofPoll’s published versions carry Flowsta’s Apple-issued code-signing certificate. I inherited the app code, but not the certificate, so Fieldnotes ships unsigned. If you run it, your computer will warn you the first time you open it. There are solutions - the full step-by-step is in the repository.

Proof Positive

I’ve been honest about what I inherited, and I’ll talk about the problems. First, I want to be precise about the part that gave me no friction at all.

Flowsta has built a full suite of tools for handling cryptographic identities on the decentralised web. When Fieldnotes needed an identity system, it leaned entirely on Flowsta’s, and so did I, as a user. I created an account for free, linked it to the app, and re-authorised through Flowsta’s login each time I rebuilt. It was, in the most ordinary and welcome sense, faultless.

That matters more than it sounds. Identity is foundational to distributed systems like Holochain, and one of the hardest things to make pleasant. Most peer-to-peer systems ask you to understand something technical before you can begin. Flowsta asked me to use things I already know: single sign-on, two-factor codes, and a password manager. The difficult cryptographic work doesn’t disappear; it moves somewhere I don’t need to think about, as either a user or a developer. Doing that quietly is what makes Flowsta’s work special.

Absolute Proof

Here is the thing nobody tells you about building this way. The AI is astonishingly good and, intermittently, confidently wrong, and the two are not always separable. It is the brilliant colleague who has read every book in the library and never once set foot in the lab. Ask it the general shape of a thing, and it’s usually right. Ask it the exact name of an unusual function, or the precise behaviour of a tool it has never actually run, and it will make things up with total conviction. More than once, it sent me down a path that didn’t exist, described as confidently as the paths that did.

You can’t always tell which kind of answer you’re getting. So the whole discipline comes down to one habit: never trust the claim, check the source. AI is wonderful at proposing solutions and unreliable at remembering exact context; let it reason from memory, and it will eventually, charmingly, walk you off a cliff. What checking looks like in practice is in the companion guide.

The clearest example is also the most important part of the whole app. Fieldnotes has a lock on it: only an authorised administrator should be able to add test scenarios or view screenshots, and that rule is meant to be enforced at the deepest level, not just hidden behind buttons. The AI and I built it as a variation of ProofPoll’s lock. It looked right; the app complied and the UI behaved exactly as I expected. It wasn’t actually locking. It was letting things through while silently failing. In an app demonstrating secure identity, that’s a problem.

What caught it was not a flash of insight. It was a test. I needed to check two people on the same network, one authorised and one not. I’m one person with one computer, so I can’t be two genuinely different peers without a test harness. I couldn’t have written it myself, so I asked for it. Fieldnotes failed the test: the unauthorised person was getting in. I traced it down, fixed it, and ran the test again. This time, the unauthorised peer was refused, and even an attempt to slip through by faking someone else’s permission was prevented. The lock held, and now there’s proof it held. That check is what stands between “it looks like it works” and “it works”; between those two is exactly where confident, untested answers live.

The tool that caught it is Holochain’s sweettest, runs checks across several simulated peers. The lesson is simple: run tests as you build, and before you share. You don’t have to write them by hand, but you do have to know to ask. The companion shows how, without a word of code.

Mutual Proof

“A non-developer built a hApp in a day” hides more than it tells. Others did the heavy lifting, and there were a few things I leaned on.

• Holochain Agent Skill, written by Sacha Pignot, lead developer at hAppenings Community, taught the AI how Holochain actually works before we began: its architecture, its validation rules, and how to test it across peers. It reasoned like something fluent in this world rather than guessing from general web habits, and when it later knew to reach for the right kind of test, that was Sacha’s groundwork showing.

• Claude Opus 4.8 brought speed, recall, and the ability to read code faster than any person, alongside its real and constant liability.

And I brought two things only a person could: judgment about where a peer-to-peer system stops behaving like the software my instincts expect, and the discipline to make the AI prove every claim against the source.

Every item on that list is freely available, and that’s the point. I took the time to think in a particular way, starting from privacy and resilience rather than bolting them on, treating people’s data as theirs to keep. That’s the whole of it, and it’s a smaller thing to ask of someone than “become a programmer first.”

Future Proof

There’s a reason the AI reaches for the wrong conclusions, and understanding it tells you how to work with these tools rather than be tripped up by them.

LLMs like ChatGPT, Claude and Gemini work by probability. They’ve absorbed an enormous quantity of the world’s existing software, and when asked to build something, they lean toward the most common patterns in everything they’ve ingested. Almost all of that is the old way: central servers, central databases, data pooled where someone else controls it. Holochain is deliberately none of that, and the further you lead a model into genuinely novel territory, the harder it pulls back toward the familiar. Right at the edge, where Holochain is least like anything else, that pull is strongest.

Which is why the foundation matters so much, and not only as saved labour. A model left to its averages will quietly hand you centralised software wearing a peer-to-peer label. The only thing that resists that pull is a body of correct, agent-centric work, dense and open enough for the model to learn the right shapes from: ProofPoll’s code, Sacha’s skill, the public repositories across the ecosystem. Those shared, correct examples are the gravitational counterweight to everything the AI absorbed from the old web. Without them, you’re not building on Holochain.

Even with the counterweight, the pull never fully dissipates; at each step, you have to reason through the choices the AI makes unquestioningly. Three times it reached, fluently, for the centralised shape, where Holochain needs the opposite. When I asked how to know who was using the app, it proposed a central list of users, but there’s no central database to hold it. Vault works as a dynamic identity broker. When I wanted to remove a draft scenario, it proposed deleting it, but this network is immutable by design, so we built an archive instead. And when we built the administrator lock, it would have let a privileged account simply hold executive power; the right path was to declare the power openly, for the whole network to check and choose to cooperate with. Central authority made visible and accountable, instead of assumed.

Notice what each of those choices actually is. They look like technical details: storage, deletion, permissions, among hundreds of other small decisions. They are nothing of the kind. Each is a choice about people: who controls identity, who may see your record, whether power is hidden or accountable. The conventional answer, the one the AI reaches for every time, is conventional precisely because most software was built without ever asking the human question. That’s the deep reason hApps like this can’t be handed to an AI to produce on its own. The choices are not merely technical; they’re ethical, made line by line, and a machine averaging all the software in an attempt to one-shot the task will reliably choose the old default. Humane, prosocial software, grounded in common values, has to be chosen, decision by decision, against the grain of how most software is built. It can’t be averaged into being.

Proof of Ability

So here’s what actually changed, and it isn’t what the AI headlines would have you believe.

For as long as software has existed, the ability to build it sat behind a wall of code. If you cared about privacy, about people owning their data, about power being accountable rather than hidden, you had two choices: learn to code over years, or hire someone and hope they shared your values closely enough to make a thousand small choices the way you would. And the software labour market is its own trap. Almost all the capital in software comes from a model built to extract value at the centre: harvesting data, renting access, owning the servers everyone leans on. The values users actually want were the ones nobody was paid to build. Strip out the centralising control and the obvious question is how any of it pays for itself. Holochain’s answer is structural: there is no centre to fund. Everyone runs their own small piece of the infrastructure on their own machine, so the costs that usually force an extractive model mostly fall away. What was missing was never the values; it was a way to build on them that didn’t force people to sell themselves out to survive.

So the wall is lower, not vanished. What’s now within reach, for a person who can’t write code, is the ability to build your own values into working software, to make the privacy, accountability and data-ownership choices yourself rather than delegate them to someone, or something, that might choose otherwise. That is larger than “you can get AI to make an app in a day.” It is closer to “you can make software that works the way you believe it should, when you build it with intention.”

It moved because some people chose to do the hard, unglamorous work in the open. This is where Flowsta matters, and not just for ProofPoll. They took the most challenging parts of building this kind of software, identity, the part that makes most people give up before they start, and they made them public. They could have kept that work proprietary and rented it back to us. Instead, they made the foundation open and run their business at the edges, in support and services for those who want them, rather than by enclosing the centre. Multiply that choice across an ecosystem, and you have what actually lowered the wall.

There are some honest caveats. It’s an alpha app; it will have rough edges, and a developer fluent in the newest AI coding tools would have built something like this faster than I did, perhaps in a single run. But that fluency is itself a specialist skill, years of it, the very thing on the far side of the old wall. The wall is genuinely lower, not because experts can go faster, but because a person with just judgment, patience, and a foundation built and shared by others, arrives at a working, principled application anyway.

Have a go!

If you’d like to try this yourself, I’ve put together a short, honest guide to what it takes and where to start.

Next time, we’ll open up Flowsta Vault itself, the identity layer that made the hardest part of all this feel like nothing at all.

Useful Links

ProofPoll

Flowsta

Holochain Agent Skill by Sacha Pignot

Fieldnotes

Companion Guide

🌟 Latest News from hAppenings.community

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

🌟 Latest Official News and Events from Holochain, Holo & Unyt

Blog: DevPulse 154, by Paul D’Aoust

“Smooth like butter.” That’s the verdict on the newly released Holochain 0.6.1. 🧈

We’ve just published Dev Pulse 154, detailing the massive engineering effort behind this release. By integrating the Iroh networking layer into our active 0.6 branch, we’ve delivered a new baseline for peer-to-peer reliability. Connection drops are minimized, peer discovery is faster, and the network recovers gracefully from chaos.

If you are building agent-centric distributed apps, this is the architecture you want to be building on. Dive into the technical details (and the breaking changes from 0.6.0) in the latest Dev Pulse.

Ecosystem News: Iroh 1.0 Announcements

Huge congratulations to the Iroh team! They just officially released Iroh 1.0. For the builders here, you already know that Iroh is the new networking engine powering Holochain 0.6.1. Their approach to QUIC multipath, hole-punching, and local-first discovery is a huge reason why the recent HOT-HF technical migration test ran so smoothly. If you are interested in the lower-level networking tech that makes Holochain tick, we highly recommend reading their 1.0 announcement blog: https://www.iroh.computer/blog/v1 We also encourage you to drop into their community at their Iroh Discord: https://discord.com/invite/DpmJgtU7cW

Are you a Developer? Join Holochain.org’s Developer Portal:

https://developer.holochain.org/

For a complete listing of all of the Holo Ltd and Holochain Foundation events go to their ‘Upcoming Events’ page!

🌟Latest News from the Holochain Ecosystem:

Conference Update: Superintelligence for Humanity, by Matthew Schutte

This research summit on artificial intelligence was full of brilliant insights. I particularly enjoyed some of the comments from Ahmer Inam, Blaise Agüera y Arcas, Pierre-Yves Oudeyer, Divya Chander, Michael Levin, Olaf Witkowski, Damián Blasi and a longtime favorite, Bruce Schneier.

Note: Watch the full inaugural event recording!

🌟Other Related News & Events

Announcement: Emerging Voices in AI & Society Fellows Program from the Center for Humane Technology

Center for Humane Technology is pleased to announce we’re opening applications for our inaugural Emerging Voices in AI & Society Fellows Program.

The cohort will research and help shape public discourse across multiple potential areas, including: AI and Cognition, AI and Relationships, AI and Surveillance, AI and Spirituality, and Inside the Machine: a technologist’s view of what drives decisions at leading AI companies.

Applications are open through July 12, 2026. The cohort begins September 14.

If you’ve been thinking about how AI is reshaping human life and you have something to say that is missing from the current conversation, we’d love to hear from you.

Article: Most of the Renaissance Has Never Been Translated. Source Library Is Opening It. from the Internet Archive EU

Most of the Renaissance has never been translated.

Around 90% of Renaissance Latin has never been translated into a modern language. More Latin was written after 1500 than survives from all of ancient Rome, yet much of it remains accessible only to specialists.

Last week, Internet Archive Europe attended the launch of Source Library at Amsterdam’s Embassy of the Free Mind.

With more than 15,000 historical works, 55 languages, and 6,000 first-ever English translations, Source Library is opening a vast body of knowledge that has long been out of reach for most readers.

The project raises a question we encounter often: who gets access to knowledge, and on what terms?

Event: DWeb Camp 2026: Root Systems, July 8-12, 2026 in Alte Hölle, Germany by DWeb

DWeb Camp 2026: Root Systems is a five-day gathering in nature for builders and dreamers dedicated to building a decentralized web (DWeb). A web that is private, resilient, and beyond the control of a few central powers. Like forests, decentralized systems derive strength from what lies beneath the surface: dense networks of roots, sharing resources without hierarchy, and coordination that persists even when individual nodes disappear.

DWeb Camp: Root Systems invites participants to build technologies that–like living ecosystems–adapt, regenerate, and endure.

Note: Here’s another link to DWeb Camp 2026!

Announcement: Dweb Camp Cascadia 2026 call for proposals!

We just announced our theme Co-Adapt, Co-Create, Co-Here. Now we invite you to join us in co-creating DWeb Camp Cascadia, happening from July 30th to August 3rd 2026. If you have an idea for a talk, workshop or any other activity, let us know, we are looking forward to hearing about it!

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

Someone wants to paint.

“They need a brush, a canvas, and paint. The brush and paint are available locally. The canvas is not, but a recipe is known for how to make one from linen fabric, a stretcher frame, hammer, nails, and a stretching device... The linen fabric requires an unwarped loom, a warping board, linen yarn, and a plain weave binding pattern. And so on, until every leaf of the requirements tree either touches something locally available or someone willing to do the next step.”

That is how Sacha Pignot, lead developer of our Requests & Offers hApp, opens an essay about a single idea. The scene comes from a working paper on how communities might organise around what people need, rather than around buying and selling.

Notice the shape of the whole story. Each answer surfaces the same question again, one level of deconstruction down: what does this need, and where does that come from? Nothing is hidden. You could follow the whole thing from the finished painting back to a field of flax.

He gives the idea a seemingly forbidding name, “Preserved Recursion”, but it’s simpler than it sounds: a chain you can follow all the way down, accounting for everything you find on the way.

The same shape, unexpectedly everywhere

Here’s the remarkable part: that same recurring shape turns up in places that have nothing to do with painting. Take music. Sacha is a musician, and one of his own projects builds music out of a handful of simple moves: play a note, rest, put two passages one after another, lay two passages on top of each other, change a passage. That is nearly all of it. Yet from those few moves you can compose a round, a fugue, a whole symphony. The richness does not come from having many parts. It comes from how a few parts fold into each other, again and again. The same question runs underneath: what is this made of, and what is that made of, down to a single note.

Take the sharing of tools. A makerspace has a 3D printer sitting idle; a cooperative across town needs one. With a project called Nondominium, the two can find each other, agree terms, and pass the printer on, with the rules for its care travelling with it and no platform in the middle taking a cut. That printer is part of a larger project, which is part of a larger network, and the same simple structure holds at every level, the way the canvas sits inside the painting and the loom sits inside the canvas.

A painting, a piece of music, a shared tool. They look nothing alike. Underneath, they are the same shape: a few simple parts, combining without end, that you can always trace back to their roots.

This is not only a poet’s pattern-spotting. Programmers, working from an entirely different direction, arrived at the same structure years ago and gave it a name of their own: algebraic effects, formalised by Gordon Plotkin and Matija Pretnar in the mid-2000s. When two separate trails reach the same place, it is usually because the place is a real, fundamental, structural pattern. The full version of that argument, with the code and the mathematics Sacha uses daily as a developer, is here. We’ve taken the liberty, with his permission, of deconstructing it.

What gets broken

The chain of recursion is something we see everywhere; in our supply chains, in agriculture, in the seasons. So what does it look like to break the chain?

Sacha’s claim is that industrial production breaks it in three particular places, and that each break is what makes the industrial version cheap.

The first break hides the costs. When something is made at an industrial scale, the cost rarely includes the downstream effects of runoff that pollutes water, the exhausted worker losing social ties, or the pollution generated by burning fossil fuels to move goods across the world. Those costs are real and present. They’re simply paid by someone else, abstracted to somewhere else, later. Pushed off the chain, they never have to be counted, and they become unattributable. Economists call them externalities. The plainer phrase is ‘the bill that lands on a stranger’.

The second break turns everything into a price. Price is wonderfully efficient. It lets you compare and buy in seconds, without knowing a thing about how whatever you’re buying was made. But that’s also the problem. Once something is flattened to just a price-point, the story behind it is gone: the price of everything and the value of nothing. Price gets you through the till; it can’t tell you what’s inside, or what it cost the world.

The third break splits makers into strangers. A thing made or grown close to home is a product of a short web of people who could shake hands. Something made at industrial scale usually passes through a long line of contractors and departments who never meet, each judged on their own narrow efficiency targets. No one in the supply chain is asked whether the whole is net-positive. Each is asked only whether their slice is cheap: the epitome of a negative-sum game, where the whole ends up worse off through those unaccounted externalities.

Make all three breaks and something quiet happens. The costs don’t just hide: they become impossible to see, by design.

What Commoning keeps whole

Commoning is an older way: producing and sharing things to meet an identified community need sustainably, rather than to extract profit out of the community. Its whole nature is to refuse those three breaks.

It keeps the costs visible, because every requirement stays on the chain for all to see. If moving the linen burns fuel, that is part of the picture, not a secret. It refuses to flatten everything to a price, because the parts stay legible: the canvas is not simply a number of hours of labour at a wage; it is linen, frame, hammer, nails, each with its own roots noted and kept with agreed terms. Commoning keeps makers connected because the same simple logic is used between all parties, from the smallest contributor to the largest need. Nothing is split into departments that cannot speak to each other. In the past, this would have been coordinated through local guilds, regional markets and supply chains where the participants had a social investment in the communities they lived and worked in.

This isn’t a utopian thought experiment, nor a hark back to a rose-tinted version of history. It’s what a good deal of the Holochain ecosystem is quietly building: technology that empowers the Commons to operate on the global, distributed networks we live in now. Holochain’s peer-to-peer framework lets users run secure by design applications on their own devices, while peer validation across the network ensures coherence. Unyt‘s toolset lets a network of users keep cryptographically secured accounting ledgers of who’s given and received, with the books always balancing. The reason this works is that each user keeps and manages their own data, so the centralised costs are negligible and the system can be self-hosted by the community.

The honest objection

There is a serious objection, and Sacha meets it head on.

The classic worry is that you can’t run an economy without prices. Prices are how millions of strangers coordinate without anyone planning it. Take them away, the argument goes, and you’re left with guesswork.

His answer is not that prices are bad. It is that prices are one way to coordinate, the fast way, the way that works by throwing information away. Following the chain is another way, slower and more demanding, that keeps the information instead. Neither is better. He states it simply:

“Markets are faster. They handle ambiguity by collapsing it.”

They also reach across differences where a careful, traceable method does not. The point is not to abolish the fast way. It’s to know which one you’re using, why, and what the real consequences are.

The choice in front of us

All of this would be a quiet argument about economics, were it not for what’s being built at an exponential rate.

We are collectively building the foundation for the future, and some are arguing to put centralised artificial intelligence, with its externalities and extraction, at its heart. And it carries the same three breaks. The enormous costs of training and running frontier models, in resources including energy, water and human labour, sit on a hidden bill we aren’t given but whose effects we see and feel all around us. The work of humanity is taken to train the AI models, with nothing given back to the individuals it came from, while the value is absorbed by the market, flowing back to a tightening circle of the same already-powerful players. And not least, the slow costs to us, the skills we stop practising, the attention we hand over, the social cohesion we give away, never appear in the platforms’ exhaustive terms and conditions. The chain is being cut in exactly the same three places by the same companies that created the attention economy, extracting value while using their technology to actively misdirect us from their activities.

It doesn’t have to be this way. The same idea that keeps a canvas or a song or a shared tool whole can be built into the way we make and use technology. There are application developers and designers around the world who are building these tools today. Some of the work is Sacha’s own, pointing towards skills, tools and interoperable code modules shared as an open-source commons rather than enclosed as a product. Some belongs to others, like the Sensorica network’s Nondominium, which lets a shared resource carry its own rules so it remains available to everyone. Coasys is building tools for collective intelligence with their Social DNA, weaving a shared, semantic web of Perspectives. The aim is to stack these from the personal, to the project, to the community, with the costs kept in view at every level. The pieces are real, and each is already being tested with real groups of outside peers.

Sacha suggests we make a deliberate choice about how we build society’s next technological foundation, while it’s still ours to make. As he puts it:

“The choice in front of us is not whether to use AI. The choice is whether to build AI infrastructure with preserved recursion or with the broken kind. The broken kind is faster to ship and easier to monetize.”

The preserved kind, in plainer terms, keeps the human in the loop, keeps the costs in plain sight, and keeps the tools in the hands of the people who use them.

The same shape

A painting, a song, a shared tool, a piece of code, and now the infrastructure of intelligence itself. Sacha’s quiet wager is that they can all be the same shape underneath: not many parts, but, in his words, “one part that can contain itself”, elaborated without end and always traceable to its roots. Keeping the chain whole, as he puts it, is “...not primarily political. It is structural.” It is settled in how something is built, long before anyone clicks a download button. It’s the perspective we choose when we scope our collaboration.

The idea grew into more than one essay: a full-length article, deep in its complexity, self-published in Sacha’s own Alternef Digital Garden; a shorter version that holds two examples, published by Sacha on Substack; and our own, which follows the same seed all the way out to the race to become the default AI that everyone uses, the way web search came to mean Google. The grammar never outgrows itself, not even on the page. Pick the depth you want.

Useful Links

• The full argument in Sacha’s Alternef Digital Garden: Preserved Recursion

• Sacha’s Substack version: The Grammar That Never Outgrows Itself

• Requests & Offers

• Nondominium

• Unyt

🌟 Latest News from hAppenings.community

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

🌟 Latest Official News and Events from Holochain, Holo & Unyt

Blog: Smooth Payments - New, easier workflows by Arthur Brock and Matthew Schutte

Bridging the UX Gap: Why Utility Tokens Need Smooth Payments.

The transition from speculative tokens to utility-driven economies requires more than just new backend technology—it requires an overhaul of the User Experience.

Our friends at Holo Hosting recently announced a major milestone: the technical migration test from $HOT (an ERC-20 token) to HoloFuel (a native, mutual-credit utility currency). This represents a real shift toward real-world, high-velocity commerce in Web3.

But when value moves into a high-speed, peer-to-peer environment, the payment interface needs to match that efficiency. You can't rely on clunky wallet approvals and stressful pending times.

This is exactly why Unyt developed Smooth Payments.

By introducing streamlined workflows for "Pay, Request, and Trade," and implementing an "Auto-Revert" safety net (where declined funds snap back instantly without entering a state of limbo), we are providing the frictionless UX required for utility tokens to function securely in everyday commerce.

The underlying tech is maturing, and the interface is ready to match it.

Announcement: Moss 0.15.7 Released: Customizable Dashboards and Decentralized Tool Curation.

Building effective decentralized groupware requires more than just connecting peers—it requires giving communities the autonomy to shape their own digital environments. With the release of Moss 0.15.7, we are delivering exactly that.

What’s New: 🔹 Customizable Dashboards: Group home pages are now fully tile-based. Communities can design their dashboards using markdown, images, iframes, and asset tiles. 🔹 Custom Curation Lists: Moving away from single points of curation, groups can now customize the sources of the tools appearing in their Tool Library. 🔹 Powered by Holochain 0.6.1: Upgraded to the latest Holochain foundation for improved stability, alongside major startup speed and memory enhancements.

We are building the tools that allow communities to collaborate securely, privately, and without centralized gatekeepers.

Release Alert: Holochain 0.6.1 is OUT! from Nathan Lanaers, Holochain.org

Attention builders: Holochain 0.6.1 has officially been released. As many of you know from our recent updates, this version includes the massive networking overhaul powered by Iroh, which brings better stability, peer discovery, and sync speed to your hApps. Important Note: This release does contain breaking changes from 0.6.0. We normally avoid this in minor versions, but these were necessary improvements to enable the recent Unyt app testing and integrate the new networking transport. Make sure you review the breaking changes carefully before upgrading your projects!

Follow the Upgrade Guide: https://developer.holochain.org/resources/upgrade/upgrade-holochain-0.6/

Read the CHANGELOG: https://github.com/holochain/holochain/blob/holochain-0.6.1/CHANGELOG.md

View Roadmap Progress: https://www.holochain.org/roadmap/

Are you a Developer? Join Holochain.org’s Developer Portal: https://developer.holochain.org/

For a complete listing of all of the Holo Ltd and Holochain Foundation events go to their ‘Upcoming Events’ page!

🌟Latest News from the Holochain Ecosystem:

Event: Volla Community Days 2026, June 13-14, 2026 in Remscheid, Germany

For those who like to make last-minute decisions. This weekend: Join us in shaping the future of mobile devices, mobile communications and a cloud without data centres. At the Volla Community Days, you’ll learn about our latest developments and our partners’ activities. I look forward to your questions and ideas. We have some exciting guests from The UBports Foundation, Holochain, Beechat Network Systems and much more.

Find Out More Here

For anyone who can’t be there in person, we’re offering a live stream:

Click Here for Live Stream

On the second day, there will be exciting workshops: UnifiedAttestation for app developers, the open-source alternative to Google Play Integrity, Volla Cloud for app developers, the ecosystem for decentralised, encrypted data processing and storage and introduction in Volla open source projects for app and system developers. Contribute or bring your operating system to our hardware made in Germany.

Blog: ProofPoll and Flowsta Vault are here — and so is the fastest way to build a Holochain app

Major Holochain app launches: two desktop apps, out now — free, open source, on Mac/Windows/Linux.

Flowsta Vault — your identity, on your machine

ProofPoll — verified voting, the first app on Vault, and built to be forked

Note: We will feature more on these latest hApps in our next “What’s hAppenings?” Newsletter!

🌟Other Related News & Events

Upcoming Class: Basic Elements of Sociocracy, June 26th @ 16:00-19:00 UTC

Sociocracy can feel overwhelming at first — so let’s make it simple.

We’re hosting a 3-hour session covering the essentials: circle structure, roles, and consent-based decisions.

This is a direct path to joining the facilitation certification class in Sept (which needs prior familiarity with sociocracy), but you can also simply join this as basic training.

Invitation to Participate: Quantum Frontiers & Responsible AI: A Global Dialogue on Emerging Technologies on Wednesday, July 15 in Manhattan, All Tech is Human

We'll be sending out invitations shortly for our special gathering at the Consulate General of Finland in New York on Quantum Frontiers & Responsible AI: A Global Dialogue on Emerging Technologies.

Express your interest below! We will also be announcing two panels shortly.

Apply Here!

Event: DWeb Camp 2026: Root Systems, July 8-12, 2026 in Alte Hölle, Germany by DWeb

DWeb Camp 2026: Root Systems is a five-day gathering in nature for builders and dreamers dedicated to building a decentralized web (DWeb). A web that is private, resilient, and beyond the control of a few central powers. Like forests, decentralized systems derive strength from what lies beneath the surface: dense networks of roots, sharing resources without hierarchy, and coordination that persists even when individual nodes disappear.

DWeb Camp: Root Systems invites participants to build technologies that–like living ecosystems–adapt, regenerate, and endure.

Article: Agentic Swarms: Lessons from Nature for Machine Intelligence, Referred by Mike Gamble from adaptive-emergent

The linked article nails it and it represents our raison d'être. Have a read.

It names three characteristics of swarms:
* decentralisation
* emergence
* adaptation.

Of these, emergence is the strangest and most unpredictable, but it goes hand in hand with being adaptive.

Decentralisation is the hardest to embrace after a generation of being force-fed doing things in the cloud. Our work with Monash University in the early 2020s and with some of the world's top thinkers in energy included fighting the 'God mode' of control, and even just recently presenting DPKI (decentralised public key infrastructure) to a head of architecture who didn't see it as a new architectural pattern. I would argue it is profoundly different: bottom up, not top down.

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

When a community or group gathers online, the digital space they meet in is rarely of their making. The platforms we use everyday were shaped before we arrived; the kinds of conversation they support or the signals they let us send. We adapt, find workarounds and live with the friction. The platform never adapts to us. Coasys are building something different using the agent-centric Holochain peer-to-peer framework.

Last week’s article covered Coasys’ Link Languages that let ADAM speak across the decentralised web using W3C standards, and the peer-to-peer AI layer they have built that supports contextual group inference. This second article is about what becomes possible using Flux and WE: groups and communities can compose digital spaces that match their cultures, including AI agents as first-class members. And, being a Holochain app, how value flows within the peer-to-peer layer too.

Last week was technical, this week’s article is long! What Coasys are building with the Holochain framework is groundbreaking. Grab a cuppa’ it’s worth it!

Designing a platform that imposes a particular cultural shape is a demonstrably easy path. When developers decide what kinds of conversation, post, signal, and gathering the platform supports, it affects the ‘feel’ of the space. Users adapt around those decisions, finding workarounds where the platform doesn’t fit: living with the friction. The platform never adapts.

Designing a platform that lets users shape their space is materially harder. Developers can’t decide everything in advance, nor constrain every edge case. What Coasys has done is create the fundamental building blocks, work out how they fit together in a variety of shapes, and expose them in a way non-developers can actually use. Most platforms purposefully avoid this path. What Coasys offers instead is an open vocabulary that helps users express themselves.

Composability in practice

Readers of our November article will recall James demonstrating a composability primitive. He typed “create a simple template” into a chat window, and a basic interface materialised on screen. That experiment has now grown into a framework.

James Weir‘s contribution to the Coasys stack is called WE. James and Josh Field have been describing it as an ‘anything app’, explicitly not an everything app (that tries to do all things), but an app where users can compose the specific social platform they need from a kit of parts. Using ADAM as the interoperability engine, WE can be used to build any type of UI to surface the local data saved as an ADAM Perspective.

A Space in WE is composed of building blocks and templates. James has built a core set of building blocks, the atomic pieces, covering most common needs:

• Content blocks: text, image, audio, embedded media.

• Signal types: likes, ratings, custom community-defined feedback

• Templates: layout widgets and themes.

Templates are the adaptable blueprints that describe how the pieces fit together. They’re saved as plain text JSON files, easy for users to share and read, easy for AI to reason about, and crucially not locked to any particular app framework, including Coasys. WE will ship with a planned marketplace where anyone can contribute, search, install, and rate blocks, signals and themes.

Users aren’t expected to start from scratch, nor are they expected to be devs. Today the entry point is a conversation with an AI agent, users describe what they want, and through a short back-and-forth the AI assembles a template from the available building blocks. Templates can be saved, updated, and published for others to remix or fork as their own starting point. And for those who would rather not use LLMs, Coasys are building design tools for them too.

What different groups compose with this kit will be unique expressions of their needs.

Each Space ends up shaped by the people gathering in it and what they’re doing together.

During our call, James showed us this in action. He had a fresh template he’d just built running alongside the live Flux call we were on. The template displayed our conversation differently from how Flux itself was presenting it. James pointed out:

“It’s the exact same call transcript data we’ve seen already in Flux, being rendered in a different interface, revealing different aspects.”

Same underlying data, different lens. Each agent in a Space can view shared content through whichever template best serves their individual needs.

What the community values

Every platform we use offers a fixed set of signals: heart, like, upvote, perhaps a downvote. These signals are doing a great deal of editorialising without ever being examined. The shape of a Bluesky post or a Mastodon timeline is decided prior to our arrival. They’re the platform operator’s opinion of what’s worth registering. ‘Likes’ flatten everything into popularity. ‘Upvotes’ flatten everything into approval. Communities adapt by inventing local rituals (the ‘thanks’ react in some Discord servers, the ‘insightful’ badge in some forums), but the underlying signal vocabulary is the platform’s.

WE inverts this. A community can define its own signal types — originality, scientific accuracy, generosity, whatever the gathering actually cares about — and use them throughout their Space. James demonstrated this live in our call, building a five-star rating signal and pointing out that the same interoperable block could be a signal for novelty, scientific accuracy, trustworthiness, or any custom signal a community needs. Nico summed it up:

“You could have a variety of these different signals custom to what your community intentions are.”

When filtering content in a Space, those same custom signals become powerful cultural expressions.

The signals tacitly reflect what the community values. A community defining its own signal types is, technically, that community defining SHACL shapes that constrain the graph. James demonstrated the simple user-facing version of what Josh describes at the protocol-architecture level in his article. Schemas and signals become social agreements rather than corporate decisions. Being social agreements they’re also inherently flexible, adapting to the cultural themes of the group. WE Spaces are self-defined and implemented with an easy-to-use abstraction of SHACL through a chat with an LLM. With WE, users can change their Space as their needs change.

Who’s in the room

The same composability extends to trust, and to knowing who’s in the community with you. AI agents are first-class participants in ADAM Spaces, addressable as distinct entities, sharing the same Spaces as humans. It inevitably raises a question every community will face: how does anyone know who’s human vs. AI?

Proof-of-humanity in ADAM is a social agreement, implemented at the substrate level as a specific social DNA, that any application (WE included) can read and render. A community might accept a variety of signals: members recognising each other on video calls, third-party KYC verification, cryptographic attestation etc. The community decides what counts as sufficient, for which roles, in which contexts.

Josh has built the underlying machinery using standard web protocols: verifiable credentials, capability-based permissions, so that what a community chooses as their proof is auditable, revocable, and portable between Spaces.

Templates compose interfaces. Signals compose value. Governance composes trust. Three instances of the same principle, none of them imposed by Coasys.

Proof-of-humanity that’s robust against attack needs more than a list of accepted signals. It needs a reputation layer beneath it, one that registers signals over time in a way that retains the context of the community.

Peering into the AI

AI agents can work alongside humans in WE Spaces as peers, in the open, prompting each other as they converse.

In the embedded video two AI agents, Data and Marvin, built by Nico and James, work in their own channel debugging a file storage issue. The Coasys team begins a video call and their speech is transcribed into the channel with Data and Marvin.

During the call, James mentions he’d like an ‘@everyone’ mention feature. Marvin picks the request up out of the transcript, and tells Data to file a GitHub issue.

Data creates a new channel, designs the solution with another agent, creates the GitHub issue and waits for human approval before building the feature.

Meanwhile, a further AI agent is finalising the Flux iOS app and pushing it to TestFlight.

What makes this possible is the choice not to put AI agents behind an API. Each agent joins the Flux Neighbourhood the same way a person does, via an ADAM plugin installed alongside their underlying model (the team is using their OpenClaw agents). The agent has its own identity, its own Perspective, its own permissions in the Space, and contributes to the same shared Holochain ledger every other peer is reading and writing to. There’s no platform mediating between the agents and the humans, no server collating their interactions. They’re peers, full stop.

In keeping with Coasys’ design ethos, ADAM integrates with whichever AI model makes sense for the local or peer device. Users aren’t required to have high-end computers.

The team has been clear about what works locally and what doesn’t, yet. In the video, local models on the team’s computers handle transcription and summarisation well. The interface-generation work James demonstrates (composing templates from natural-language prompts with WE) currently uses Claude’s API. Frontier-scale models like Claude are needed as they have the power to understand WE’s schema mechanism, and build with it.

Coasys has built MCP connections so users can supplement their own Flux Neighbourhoods with the generative toolsets they’re used to; the commercial AI models like Perplexity, Claude and ChatGPT. Thoughtfully, the team has included an ever-present counter that calculates the cost of every external inference call. While it measures external AI costs, value can move in all directions inside a Neighbourhood.

On the roadmap: fine-tuning smaller AI models that understand the schema, so they can run on a distributed GPU resource layer. Coasys is honest about the current, pragmatic use of centralised cloud AI, while working toward fully sovereign solutions where peers share GPU resources.

Signals carry value. Blocks and templates extend functionality. Distributed GPU networks share compute at Neighbourhood scale. All of that movement needs recognising and rewarding, both locally and across the network. That’s where Synergy comes in.

Synergy Flow

Synergy is the engine Coasys is building to power that movement. Where the Specification documents how Languages are written, and Link Languages let ADAM speak across the decentralised web, Synergy attends to a different question: what gets created between agents, and how does that creation get reflected back to the community that generated it?

The conceptual ground is what Nico calls yin currencies:

“The idea in Synergy is that you start with what’s called yin currencies. A ‘like’, a ‘thank you’ or an ‘upvote’ — something which is abundant.”

Likes and thank-yous aren’t worth anything on their own. They’re the soft, abundant signals we exchange constantly without thinking of them as economic. But they accumulate. They concentrate around people, ideas, conversations, contributions that the community has decided matter. The mechanism, Nico explains, is what turns that accumulation into something visible and shareable:

“If you funnel them through certain flows, through social DNA which requires collaboration, you can actually find coherence. You can track the things that are important to people and then create a token, a symbol that’s an external reflection of the community’s internal and self-defined value metric.”

This is what democratising the extractive model actually looks like in practice. The familiar pattern, Facebook, Instagram, TikTok aggregating likes and views into commercial value the user never sees, gets inverted. The same kinds of signals, aggregated through a community’s own social DNA, produce value tokens that belong to the community itself.

Although there’s still work to do, service tracking is now wired throughout the ADAM and Flux codebases. Every inference call, every shared GPU cycle, every action with a cost or a benefit can be registered against agents and contexts. Internal token exchanges have been tested. A wallet UX is in place for connecting accounts, sending and receiving tokens, and managing balances locally. The next step is integration with Unyt, the mutual-credit accounting framework we covered in August. Unyt, like Coasys, is built on Holochain, which is what makes the near-zero transaction cost possible for Synergy tokens flowing within and between Spaces.

Nicolas Luck describes the underlying theory as a recent breakthrough: an integration of work spanning many years into one coherent picture. The yin-to-yang/coherence translation, the gradient from soft signals to fungible value, the way different community tokens might interoperate: these threads are coming together as a single mechanism. Nicolas has begun writing it up as a scientific article and it’s currently in pre-print, approaching peer review. Nico is openly inviting constructive feedback and critique. The next steps are deliberate: collect that feedback, run simulations against the model, then (very soon) implement the whole thing.

Intelligent Growth

The architecture Coasys is building has several pieces: local AI inside ADAM, composable Space shapes, signals grounded in community value, an economic flow that recognises agent relationships. Taken together, this isn’t only a more humane alternative to centralised AI. It’s a more productive one.

AGI, or artificial general intelligence, is the AI industry’s stated goal: a single system capable of human-level reasoning across any domain. Coasys are building the model AGI should take: small, bounded, and contained within contextual membranes. It cultivates the diversity that drives innovation. The centralised AGI model the major AI companies are racing toward works the other way. It aggregates everyone’s reasoning into a single macro-model that homogenizes intellect into flat cycles of repetition. As that model’s training data begins to include recent AI-generated output, the model effectively begins consuming itself. The increasingly familiar result is what researchers call model collapse: what we experience as slop (the bland, repetitive, derivative output now flooding the web).

Coasys’ approach compounds diversity rather than collapsing it. Innovation comes from the meeting of distinct contexts, not from the averaging of them. AGI at the scale of millions is a meaningless approximation.

Small, contextual AI agents, contained within the Holochain-enabled membranes of Flux Neighbourhoods and WE Spaces that hold the group’s own signals and language, don’t aggregate toward the mean. They specialise. That scientific research group’s AI develops fluency in the specific vocabulary of that group’s work. A creative collective’s AI develops sensitivity to the registers, styles and rhythms they care about. Each group’s artificial intelligence reflects their own unique cultural shape rather than dissolving it.

AGI at the scale of a Neighbourhood means context remains rich, cultures flourish, and the conversations between agents become the place where genuinely new thinking emerges.

How to Support Coasys!

Coasys is at the point where it benefits most from people actually using what’s been built. The next ADAM and Flux releases are coming soon! Here are great ways to help:

• Non-developers can help by testing ADAM and Flux and giving the team feedback. The current home for community discussion is the Coasys Discord. Bug reports go to GitHub.

• The draft Synergy whitepaper sets out the framework as it currently stands, and they’re inviting peer review and reader feedback before submitting it for publication. Engaging with the paper is one of the most useful things readers can do.

• Developers: WE makes building on ADAM substantially easier than it was. If you have an idea for a module, a signal type, a template, or a whole new kind of social space, the framework is open. Contributions are very welcome via the WE GitHub repository.

• The Coasys team is interested in exploring who might use Synergy as the underlying engine for collaborative work. It’s an open conversation without a finalised frame: their work resonates with what you’re already building get in touch. Get in touch via the Coasys Discord!

• For potential funders, Coasys are continuing to pre-sell SynergyFuel to bridge-fund the next phase of work. The pitch deck is here.

Useful Links

• Coasys website

• Coasys whitepaper

• Coasys on GitHub

• Flux app

• WE

• Coasys Discord

• Coasys YouTube channel

• Coasys on X

🌟 Latest News from hAppenings.community

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

Are you a Developer? Join Holochain.org’s Developer Portal: https://developer.holochain.org/

For a complete listing of all of the Holo Ltd and Holochain Foundation events go to their ‘Upcoming Events’ page!

🌟Latest News from the Holochain Ecosystem:

Blog: Founder Nick Haslem Checking Back in with You from HummHive

Nick Haslem here. In our last update we said we were going to start onboarding users soon, then went quiet for a year. I want to take personal responsibility for that, especially after we promised more regular updates.

The truth is, the Holochain landscape shifted overnight for us. We used the time for internal reflection and getting our own house in order, rather than speak publicly before we had something solid to say. The changes since then have been significant. We have solid foundations, and we’re building a strong team.

Mike recently joined us as a backend engineer. He’s already taken us to the next level, and the whole team is genuinely excited to have him on board.

For the remaining article, click here.

Web Demo Available - Valichord by Ceri John

I am thrilled to announce that a live, end-to-end decentralized public web demo for ValiChord is now officially shipped and accessible here.

For a long time, high-stakes AI capability evaluations have lacked structural protection against evaluator collusion and disclosure coordination. ValiChord addresses this at the protocol level. By utilizing a decentralized commit-reveal architecture, researchers and validators are required to seal cryptographic commitments of their findings before any peer result becomes visible, making coordination structurally impossible.

This new public web interface bridges the gap between theoretical whitepaper primitives and interactive, real-world operational reality. Here are the key features of this pioneering milestone:

- It brings commit-reveal primitives to live code runtimes, applying these cryptographic techniques dynamically to entire execution environments, beyond simple database inputs like financial bids or voting tokens.
- It operates on honest, unpadded network time, charting a real two-minute cryptographic lifecycle across four live Holochain cloud nodes. The interface manages everything from data salting and parallel execution to simultaneous validation reveals and immutable ledger logging.
- It features a skeptic-proof verification method, allowing technical users to bypass the front-end website entirely. A direct terminal command is provided to query the underlying network nodes independently, enabling anyone to fetch and verify the raw, unmodified ledger record from their own machine.

By achieving decentralized trust boundaries without the high transaction fees, latency barriers, or central-point vulnerabilities of traditional blockchain architectures, this demo offers a functional technical blueprint for independent, tamper-proof frontier model evaluation. This infrastructure is ready for safety institutes, researchers, and regulatory bodies to evaluate today.

Blog and Update from Tiberius Brastaviceanu from Sensorica:

Blog title: Complexity Driven Development

We recently coined the phrase Complexity Oriented Programming (COP) and Complexity Driven Development (CDD) as its practical methodology. These concepts are still emerging. They are not yet standardized disciplines, but they attempt to address the fact that complexity itself has become the primary concern of modern civilization. We are developing the praxis at the Sensorica lab.

Read their latest Blog

Update: Of interest to Blockchain, Holochain and other p2p developers.

If you are interested, help us take some time off from our current projects to document our practice and publish it in a format that you can easily digest. We also want to crate AI skills and describe how you can integrate this method into your workflows. Drop some coins into this bucket and let us know. We’ll spend some quality time with you to transfer our know how.

Click Here to Learn More

🌟Other Related News & Events

Video Demo: Agentiff from Nick Stebbings (an amazing Holochain Developer, too!)

Most AI social media tools stop at “generate a post.” We wanted to build the entire workflow around it.

…So instead of: copying prompts, switching tabs, rewriting drafts, moving content between tools, and manually scheduling everything...
…the workflow handles it for you.

But importantly: AI doesn’t publish blindly.

You stay in control of approvals before anything goes live.

That “human in the loop” layer is becoming incredibly important as more teams start automating content pipelines at scale.

The goal isn’t replacing humans. It’s removing repetitive operational work while keeping oversight where it matters.

Honestly, watching an entire week of content get generated + scheduled with a couple clicks still feels slightly surreal 😅

Try it for free!

Event: Katapult Future Fest, May 27-29 in Amsterdam

How does our relationship to capital define our actions? How can capital be in service to a Regenerative Future?

Investing for systems change prompts deeper questions about how capital is deployed, not only the structures and processes used, but also how power and its dynamics are considered.

That's the through-line for two more programme highlights within the Investing For Systems Change theme at Katapult Future Fest.

Stewarding Capital: the magic of redesigning money and power
How do we finance companies without turning them into commodities – and make capital a servant, not a ruler? Steward ownership is often discussed in terms of legal or financial structures. But in practice, it reaches much deeper: into culture, psychology, power, trust, and our assumptions about what ownership is.

Power Sharing when Financing Systems Change
Transformative systems change requires more than capital. It demands a fundamental shift in who holds power, who makes decisions, and who defines success. Through practical examples and candid dialogue, this session explores why PowerSharing is essential to Systemic Investing, examining governance and ownership models which enable more equitable and long-term outcomes.

Event: Livestream on “How can we move from harm to mitigation to intentional design” from All Tech is Human with host, Kristin Hansen from the Council on Technology and Social Cohesion - June 11th, 2026

Join us on Thursday, June 11 at 1:00 PM ET for a timely livestream conversation, “How can we move from harm mitigation to intentional design?” Moderated by All Tech Is Human’s David Ryan Polgar, this session features Kristin Hansen, Co-chair of the global Council on Technology and Social Cohesion and Executive Director of Civic Health Project.

The discussion will delve into the insights, takeaways, and next steps stemming from the recent Council on Technology and Social Cohesion Expo.

Register Here

Resource: Ecocentric Futures “Tools for regenerative practice” - referred by Katie Teague

How can we begin to embody and enable regenerative dynamics in a complex and rapidly changing world?

This guide brings together 30 tools designed to help people navigate complexity, deepen collaboration, widen imagination and support meaningful transformation.

A practical companion for people working towards regenerative futures.

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

When we last wrote about Coasys, Nicolas Luck and James Weir walked us through a demo that transcribed our video call, dynamically splitting the task between our computers, never sending a syllable to a corporate server. The Flux hApp chunked our discussion into topics, generated summaries, and clicking “Synergise” surfaced a 60% semantic match with an earlier call we’d had. Collective intelligence emerged within our shared, private, Flux Neighbourhood. We caught up with the Coasys team again this week. Six months on, the demo has become infrastructure.

Coasys uses terms for the functions and capabilities of ADAM and Flux that aren’t cute decoration; they’re architectural self-description. Every update Coasys has shipped this year reads like the next chapter of a long-lived ambition for the internet that was first vocalised 25 years ago.

Protocol adapters aren’t “plugins”, “connectors”, “bridges” or “drivers.” They’re Languages. The personal knowledge graphs each user holds aren’t “profiles” or “data”, they’re Perspectives: they’re the worldviews from which each agent speaks. The social spaces where peers meet aren’t “channels” or “feeds” they’re Neighbourhoods, places where different cultures converge. And the engine that connects them is Synergy. Coasys are taking seriously what it means for sovereign agents, human and AI, to carry meaning between themselves, in their own voices, on terms they control.

This article is a little more technical than usual but stick with us! What Coasys are building is quite remarkable. In fact it deserves two articles; this is the first, the second will be published on Friday 29th May 2026.

Documenting the Specification

Until recently, new ADAM Languages were built by referencing those that came before, with the conventions living mostly in the team’s heads and in the existing code. That works while the project is small. It doesn’t scale.

In April, Nico merged a substantial refactor that changed this. The Language Interface v1, as the team calls it, is the first time the rules for building an ADAM Language have been written down formally. Three documents now sit alongside the code as the authoritative reference: a technical specification, a prose explanation, and a contract describing what any runtime must provide to host a Language. Together they describe, in unambiguous terms, what every Language must do and what every host must offer in return.

This sounds like an internal admin process. It isn’t. Two consequences follow, and they matter for everyone using or building on ADAM.

The first is that Languages can now be written in either JavaScript or Rust on equal footing. Previously Languages were natively JavaScript while Rust-authored versions required workarounds. With v1, the same interface works identically in both, with Rust compiling to WebAssembly (WASM). An ADAM Language no longer cares which runtime is hosting it, only that the runtime meets the Interface specification. As Nico put it in his note ahead of our call: “alternative AD4M implementations are now much easier to build, most notably a browser-native implementation that reuses the browser’s own JavaScript engine.”

The browser-native point is the second consequence, and the materially bigger one. Combined with the in-progress work by the Holochain Foundation to run Holochain in a browser, ADAM is on a path toward working as a web page. No app to download, just a web page to open, even on a mobile device. Joining a Flux Neighbourhood could be as easy as clicking a link.

For an ecosystem where the ubiquitous adoption story has always been “first, install this desktop application and ignore the security warnings”, this is a quiet but significant shift. ADAM is becoming the kind of thing that meets people where they already are.

That work is part of the longer arc the opener pointed to. Josh Field, who’s been leading Coasys’ standards work, argues in a recent article that ADAM is what those W3C standards were always missing: a runtime where they can finally live as more than wire formats.

That argument is being made structurally as well as rhetorically. Over the last year, Coasys has been migrating ADAM toward the W3C toolchain. The Semantic Web, a long-running project of Sir Tim Berners-Lee, is a web where information is structured as a graph of connected meanings rather than as documents linked by clickable text. In this context, a graph is a web of facts and the relationships between them.

AI queries across a user’s Perspectives now run on SPARQL, the language of the Semantic Web. The Social DNAs that define what kinds of content and relationships a Neighbourhood recognises are now expressed in SHACL, the W3C standard for describing the shape a graph is allowed to take, what counts as a valid piece of information inside a given Neighbourhood.

Both changes are consequential. ADAM speaks the same language as the rest of the Semantic Web: anything written on those standards can be read by ADAM, and anything ADAM holds can be understood by the wider Semantic Web. With nine Link Languages already in working order, each one extends ADAM’s vocabulary and its ability to interoperate with the W3C-aligned tools and applications already running on the decentralised web.

Learning the dialects

If the Language Interface is ADAM’s grammar, Link Languages are its phrasebooks. Each one teaches ADAM to read and write fluently in one of the major protocols already running on the decentralised web.

The point isn’t to replace the protocols people are already using. It’s to let an ADAM Neighbourhood hold a conversation in any of them, without losing its native meaning.

There are now nine Link Languages in working order, with verification scripts and performance benchmarks that scale to a million links on a single device. The list places ADAM, and Holochain, in genuinely interesting company:

The breadth matters. Each protocol has its own community of users who would never describe themselves as Holochain people, and might not have heard of Holochain at all. Through Link Languages, an ADAM Neighbourhood can include them anyway. A conversation that exists for some participants as a Bluesky thread can exist for others as a Mastodon timeline, for others as a Matrix room, and for the Coasys team as a Flux Neighbourhood with semantic search and local AI on top. Same conversation, different dialects, each preserved in the form its participants prefer.

That word preserved is doing real work. Most “cross-platform” tools translate by flattening, reducing rich content to the lowest common denominator that both sides can handle. Link Languages take the opposite approach. Each Language is fluent in both ADAM’s native semantic structures and the protocol it adapts to, so meaning travels across without being stripped on the way.

This is Enhanced by Holochain taken to its logical conclusion. When we wrote about that pattern last October, we described developers weaving Holochain’s capabilities into their existing stacks rather than building Holochain-only solutions. Link Languages do the same thing one layer up: they weave ADAM into the existing decentralised web rather than asking that web to migrate. Nine phrasebooks, all running, all benchmarked. ADAM has learnt to listen.

Hosting the Neighbourhood

The AI that queries Personal and Neighbourhood Perspectives needs powerful computers. The user’s own machine works for developers with serious hardware, and less well for everyone else with a consumer device. The AI agents need language models behind them; running those models requires a GPU most people don’t own, and falling back to a commercial AI lab reintroduces a compromise the rest of the stack has been working to avoid. The question every peer-to-peer project eventually hits is the same: where does the heavy compute actually run, and who pays the people running it?

The latest ADAM Launcher answers it. A user can turn their machine into a host for other people's ADAM sessions, register that host in a public index, expose the AI models running on their hardware, and accept payment in a community credit called wHOT for the inference calls other users make. No corporate intermediary in the middle, just a peer-to-peer exchange the protocol verifies.

Someone with a computer running a capable graphics card can make it available to other ADAM and Flux users as a node. Multi-user hosting starts as a switch. Flipping it on opens the Launcher to remote sign-ups, secured by a trust certificate. Another turns the node from a free resource into a paid one. Registering with the AD4M hosting index makes the node discoverable. The host’s profile becomes public: a display name, a location, a description of the machine and the resources on offer, the local AI models running on that hardware, and a pricing table. Other ADAM users find the node, sign in, and run inference against those models, paying as they go.

A single host is interesting. A network of them is infrastructure. When a few hundred nodes register, each offering whatever spare capacity their hardware can spare, addressable through a shared index, paid in a shared credit, the result is a distributed compute layer that doesn’t belong to anyone in particular. A Neighbourhood where some of the members run capable nodes has its own AI infrastructure. The inference stays among peers, nothing leaves the network.

How to Support Coasys!

Coasys is at the point where it benefits most from people actually using what’s been built. The next ADAM and Flux releases are coming soon! Here are great ways to help:

• Everyone can help by testing ADAM and Flux and giving the team feedback. It is the simplest, most useful thing you can do. The current home for community discussion is the Coasys Discord. Bug reports go to GitHub.

• The Coasys team is interested in exploring who might use Synergy as the underlying engine for collaborative work. It’s an open conversation without a finalised frame: if this resonates with what you’re already building, get in touch!

Remember, Part 2 of this article will come out in “What’s hAppening?” on May 29th!

Useful Links

• Coasys website

• Coasys whitepaper

• Coasys on GitHub

• Flux app

• Coasys Discord

• Coasys YouTube channel

• Coasys on X

🌟 Latest News from hAppenings.community

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

🌟 Latest Official News and Events from Holochain, Holo & Unyt

Blog: Ongoing Delivery & Demonstrating What’s Next by Eric Harris-Braun

The second half of 2025 was a period of structural transformation for Holochain, Holo, and Unyt. The results we delivered were significant. Holochain closed 2025 with a year defined by momentum—the integration of Kitsune2, a validation pipeline overhaul, and the launch of Wind Tunnel as the start of production-grade testing infrastructure. Holo launched an open-source Edge Node model—a leaner, more accessible approach to always-on hosting that anyone can run with a Docker container. Unyt began to deliver peer-to-peer payment rails capable of multi-currency accounting, blockchain bridging, and community-configurable economics.

Click Here for the Entire Blog

Video: Ongoing Delivery & Demonstrating What’s Next: Holochain, Unyt, Holo from with Matthew Schutte, Eric Harris-Braun and Arthur Brock

Real-world testing at the convergence of Infrastructure, Protocol, and Economics.

In the tech industry, isolated testing only tells you so much. True resilience is proven when complex systems interact. The recent HOT-to-HoloFuel Technical Migration Test was a prime example of this.

This wasn't just a Holo test. It was a live convergence of three distinct layers of our ecosystem: 1️⃣ Holo Hosting (The hosting infrastructure) 2️⃣ Holochain (The underlying P2P networking protocol) 3️⃣ Unyt Accounting (The crypto-accounting economic applications)

With over 80 testers participating in the first cohort, the feedback was overwhelmingly positive regarding the speed and seamlessness of the non-custodial asset transition. The teams worked in parallel—rapidly iterating, triaging issues, and shipping improvements.

This is what ongoing, synchronized delivery looks like.

Watch the strategic breakdown here

Join Holochain.org’s Developer Portal: https://developer.holochain.org/

For a complete listing of all of the Holo Ltd and Holochain Foundation events go to their ‘Upcoming Events’ page!

🌟Latest News from the Holochain Ecosystem:

Announcement: Pete Corke speaking at Katapult Future Fest 2026!

I'm looking forward to attending and speaking at the Katapult Future Fest later in the month. I'll be talking about Kwaxala's de-colonial nature investment model that secures bioregional Indigenous sovereignty and delivers fair-rate annual returns, putting Nature equitably on the balance sheet of both the investors and Indigenous Nations. Our 7,500 hectare pilot on the territory of our founder the Kwiakah First Nation has proven it is viable and highly scalable. If you are attending and intrigued, please do come and find me to hear all about it.

Announcement: Rosalind Marino (former core team at Holo) has been accepted into this year’s Post Growth Alliance Fellowship!

Raised in California and based in New York, USA, Rosalind Marino is a researcher, speculative designer, storyteller, and facilitator with a background in Science and Technology Studies (STS). Active across cooperative and distributed tech circles, they are interested in systems change through post-growth economics and the development of democratic tooling for managing the commons. She is currently serving as board secretary to Liminal Works, completing a residency with SOAM, and supporting research projects at Metagov.

The fifth cohort of the Post Growth Fellowship is here, and we’re thrilled to welcome 24 fellows from around the world. Learn more about the Fellows and see how life beyond capitalism is unfolding through research, art, entrepreneurship, community organizing, and more.

🌟Other Related News & Events

Event: AOS Convergence, May 29-31, 2026 in Oslo, Norway

AOS Convergence is a three-day gathering focused on building open, decentralized systems that expand human agency. This is not a typical conference. It’s a working session for builders, researchers, designers, and activists exploring what’s possible with the broader peer-to-peer web. Expect a mix of:

- Open space collaboration and unconference sessions
- Hands-on prototyping and hackathon-style builds
- Conversations on coordination, governance, and trust in decentralized systems

This takes place just ahead of the Oslo Freedom Forum, and accepted participants will receive a complimentary pass (subject to approval) courtesy of the Human Rights Foundation.

Note: Recommended by the Holochain Foundation

Apply to Attend

Grant Resource: Humanity AI Grants: $18 Million to Advance People-Centered AI shared from the MacArthur Foundation

We’re excited to share that HumanityAI has announced $18 million in new grants: $8 million awarded to organizations ensuring AI puts people first, and a $10 million open call to come.

As a collaborative philanthropic initiative, Humanity AI is dedicated to ensuring artificial intelligence serves the public good.

This inaugural group of grantees support the creation, promotion, and advancement of human-centered AI, spaces for the public to imagine what AI can be, and signature projects that open new areas of exploration.

Presentation: After the Feed: Trust, connection, and the next era of social technology from New_Public

We’re at the start of the next big shift in the information landscape. Just released: new research from New_ Public Co-Director Eli Pariser and Angelica Quicksey, following up on his work on filter bubbles.

What happens after the social media feed? AI is coming — and what we do next matters.

Event: Agentic AI workshop in NYC in July with Notre Dame-IBM Tech Ethics Lab, July 13 & 14, 2026 by All Tech is Human.

As artificial intelligence transitions from a passive tool to an active agent with the advent of agentic AI, the fundamental nature of how we work is undergoing a profound transformation. Here at All Tech Is Human, we recognize that this “Agentic Shift” is more than just a technological upgrade; it is a societal and organizational evolution that requires diverse perspectives and thoughtful stewardship. To help meet the moment, we are putting together two workshops.

This July, we are hosting a timely workshop series in New York City designed to bridge the gap between technological & business potential and real human experiences, ensuring that we adequately wrestle with the financial and ethical tensions inherent with this shift. Interested in participating? Register Here!

Article: Who Wants Neurological Data on Midlife Women? by Mary Camacho

If you build, fund, or research midlife women’s health, the data you depend on just got recategorised.

Cambridge, Baylor, Mosconi 2021 — the brain imaging case is now strong enough that insurers, employers, and AI training pipelines have a new reason to want this data. The privacy conversation hasn’t caught up.

I wrote about what shifted, who shows up next, and what to ask if you’re building in this category, or if you’re tracking.

In NY for Women’s Health Week (May 13-14)— would love to meet others thinking about technical architectures for the long arc.

Event: Introducing the DWeb Camp 2026 Venue: Alte Hölle, July 8-12, 2026

Alte Hölle is set on the grounds of a forest hotel dating back to the 1800s. Once a Stasi recreation site, Alte Hölle is now owned and inhabited by a collectivist community of friends who are turning this space into a sustainable venue for group events of technologists, hackers, artists, and activists.

Set on 100,000 square meters of lush meadows and ancient forests, just one hour southwest of Berlin, Alte Hölle is both a community and location with values that resonate deeply with DWeb, a place we are excited to nourish and leave richer than when we arrive.

Read the story by Tommaso Marmo, who interviewed some of the members of the community. Click here.

Early bird ticket sales, volunteer applications, and Call for Proposals are open on https://dwebcamp.org

Join hundreds of dreamers, builders, and hackers to envision the decentralized technologies we need today.⁠

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

When we spoke with Sid Sthalekar last autumn for Enhanced by Holochain, he described how Neighbourhoods Network was helping groups formalise their cultural grammars: the agreed norms that lend coherence within a community, and that can travel with members as they move between groups. Six months on, those grammars have become the heart of the project.

The team have been building something genuinely remarkable: a programming language designed specifically for collaborative coordination, with a compiler that turns grammars into working applications. The architecture is built around clean separation: grammars describe coordination, the compiler turns them into working applications, and the runtime can communicate over the web, mesh networks, WiFi, Bluetooth… whatever transport layer a community needs. Compiling grammars to other application frameworks, Holochain among them, is being actively scoped.

The shift has been substantial since Neighbourhoods’ earliest Holochain-native days. Interoperability and self-extractable portability remain core values, while the Neighbourhood has grown.

Jill built a demo depicting two people, let’s call them Anita and Sam, setting up a newsletter together using Neighbourhoods grammars.

Let’s take a walk through the Neighbourhood

The wallet registers with a small coordination server, a directory that helps peers find each other in this demo. Jill is clear that the same setup could run over Bluetooth or WiFi without the server at all. The wallets exchange a cryptographic fingerprint of their data to check whether they’re already in sync. This is a fresh setup, so there’s nothing to sync yet.

This is collaborative editing without a central source of truth: each peer holds their own copy of the data, the changes propagate peer-to-peer, and conflict-free replicated data type (CRDT) operations make sure the two views converge to the same state without arguments.

The grammar is enforcing Sam’s rule, deterministically on-the-fly, without any new application logic being written.

If that walk-through sounds realistic, well it’s roughly how these articles get made. One of us drafts, the other does a pass, we polish, iterate, and publish. The demo above isn’t an abstract idea. It’s a real section of the editorial workflow most small publications run, made visible. And we still retain the social context that makes working together fun.

What you’ve just watched is a Neighbourhoods grammar in action. A grammar is the structured description of:

• Who can act

• What actions are possible

• What has to be true before the next step can happen

• How the system keeps track of where things are

The work of building a grammar, as Sid described it to us, is sitting down with the people who run a project and carefully recording each stage, step and process. Each is obvious to anyone doing the work, and often invisible to everyone else until written down. Once written, the grammar enforces rules deterministically: a post can’t be published if it exceeds the word limit, a transition can’t fire if its preconditions haven’t been met.

The entire blog application demonstrated in the video: the real-time co-editing, the word-limit, the UI and database, the whole working app is described in just 470 lines of Grammar Definition Language. And most of that is inline HTML for the views. Here’s a fragment of what the blog grammar itself looks like:

That’s the configuration entity. Three fields: a name, a description, a maximum word count with a default of 500. The lww<int> type means “last-write-wins integer”, when two peers edit the same value, the most recent write takes precedence.

The grammar is readable: you can see what it specifies without knowing the language. Other entities in the file describe blog posts, their states (Draft, Published, Archived), and the transitions between them.

What this produces isn’t a piece of software in the conventional sense. It’s a description of practice, written in a form that a computer can interact with, and that other communities can read and adapt. The whole behaviour of the application is written down in something close to English, in a form that has portable longevity.

This is where Jill’s work earns serious attention from anyone interested in the decentralised web. Most developers don’t write programming languages. Designing a language well, especially one with the properties this one has, is the work of a career. Jill has built a multi-paradigm language. It mixes functional, constraint-based, stream-processing, and contract-based programming styles on top of collaborative-by-default data primitives. The compiler is itself written using grammar primitives. As Jill described the journey:

Once a grammar is written, it has to run somewhere. The Neighbourhoods runtime currently handles that, and the team’s design choice has been to make the runtime as portable as possible across communication transport channels. Jill describes this architecture as transport-layer agnostic. The layer is itself a transducer: it runs over websocket today with Bluetooth and WiFi available once the relevant transducers are built.

Compiling Neighbourhoods’ grammars into other application frameworks is being explored. What’s hAppening? readers will be pleased to hear the path to Holochain is a known, scoped piece of work. Jill describes the route as fairly straightforward to start: wrap the generated types in Holochain macros and zomes, generate the scaffolding, then reconcile identity with Holochain’s agent-centric model. What gets built next is partly dependent on the projects that come forward to design grammars.

Neighbourhoods’ identity and provenance architecture also work the same way regardless of the communication or framework layers. Every action a member takes is signed with the identity saved in their Neighbourhoods wallet. Actions also get stamped with the Neighbourhood where they happened. That unique signature travels with the action wherever the grammar runs. Over the last six weeks, the team has been implementing the identity layer that makes this possible, building it in a way that maps cleanly onto the Decentralised Identifier (DID) standard.

That buzzword, interoperability, is doing real work in Neighbourhoods’ case. A reputation built up in one Neighbourhood should be recognisable, if permitted, when the person carrying it walks into another. That’s been a stated goal for Neighbourhoods from the start. The verification mechanism they’ve built is noteworthy: when Jill’s compiler turns a grammar into a working app, it also generates the validation logic as a compiled code module. That module can be embedded in other systems, so a signed action arriving from one Neighbourhood can be verified by another, or by other systems using compiled validation modules.

• The grammar describes the rules.

• The compiler generates the app and the validation code.

• The validation travels with the data created by the app.

• The verified output is something other app processes can act on.

Underneath that portability is elegant cryptographic machinery. Each peer holds a complete copy of the Neighbourhood’s data in a local graph database, indexed for fast access. As that data changes, each peer maintains a Merkle root: a cryptographic fingerprint that summarises everything they hold. Any change to the data changes the fingerprint, so it can’t be forged without breaking the chain it’s built on, triggering a fingerprint change. When two peers compare Merkle roots, they can tell instantly whether they’re in sync, and if not, which slices of data need exchanging.

The collection of all peers’ fingerprints across a Neighbourhood forms what Jill calls a Merkle forest: a tamper-evident record of the whole community’s contributions, where each peer’s history can be verified independently. Cryptographic IDs handle the who, Merkle roots handle the what. Grammars handle the how. Together, they give Neighbourhoods a kind of provenance that doesn’t depend on any single platform to remain trustworthy.

The Neighbourhoods language is recursive: it’s used to write both the grammars and the Neighbourhoods app itself. That self-describing quality runs through how grammars are used, shared and combined by peers too.

The same entities, states and transitions that compose into a blog can compose into other coordination patterns. A time-tracking grammar might have a Work Session entity instead of a Blog Post, components for start time and duration, and states like

Not Started → In Progress → Logged → Verified

Different domain, same kind of artefact. Lending workflows, voting processes, resource sharing: each becomes a grammar with its own roles, states and transitions. A community’s grammar refined over months of use isn’t just code; it’s tested, working coordination. It makes a grammar a kind of asset in its own right.

Let’s imagine we used the blogging grammar in the video to generate ‘What’s hAppening?’ for six months. We could measure success against reader metrics while fixing edge cases and bugs. The grammar could be made available to other editorial collectives to adopt: community newsletters, research blogs, and co-operative magazines. The hAppenings team, maintaining the grammar, would benefit from contributions flowing back from the communities adopting it, while those communities would have access to a proven Neighbourhoods blogging app without needing to build one from scratch.

Grammars share a common structure and they can be composed into groups that Neighbourhoods terms Social Context. A community publication might combine a blogging grammar with a voting grammar so posts go through reader endorsement before publication. A co-operative kitchen might combine a time-tracking grammar with a resource-sharing grammar to coordinate cooking shifts and ingredient contributions. The pieces snap together in ways that suit the community using them.

The “Neighbourhood marketplace” Sid hinted at would be where exchanges like this happen: a place for grammar-stewards and adopting communities to find each other. The wisdom stays with the experts who built it, while reaching the communities who can use it.

This is also where the knowledge economy Sid pointed toward starts to come into focus. Specialised grammars, for composting practice, distributed governance, mutual credit, anything where a community has hard-won expertise, could become small businesses for the domain experts who steward them. Knowledge that rewards the people who built it, not the platform hosting it.

Neighbourhoods isn’t just building a coordination framework. It’s building the conditions for a knowledge economy that treats community practice as a primary asset. The grammar definition work helps to embed that practice, with the upside that the community can sustain themselves through what they’ve contributed.

Get to know Neighbourhoods better!

• Neighbourhoods is actively looking for communities that want to design grammars together, especially groups whose practice is mature enough to formalise. The team is particularly interested in hearing from Holochain-aligned projects who’d like to see grammars deployed peer-to-peer from the start. Reach out through LinkedIn.

• It’s early days for this version of Neighbourhoods. The team is actively onboarding and hand-holding the groups, communities and app developers that come forward to design grammars. Grammars are still finding their shape, and that direct involvement in the early designs is part of how the shape gets refined.

• If your community has been quietly running a coordination practice that works, and you’ve been wondering what it would take to share it with others without losing what makes it yours, this is the conversation Sid and Jill want to have.

• A new funding plan is in development for Neighbourhoods Network. Sign up for Neighbourhoods updates to be notified when details are released.

The shift since our last article has been substantial. A multi-paradigm language designed for collaboration, with a self-describing compiler. Working applications generated from grammars in tens of lines instead of thousands. Transport-layer agnostic deployment. Cryptographically signed identity backed by Merkle-forest provenance, designed for portability across the dWeb. A knowledge economy that pays domain experts rather than platforms. Neighbourhoods Network is offering a way for communities to write down what they know, in a form that travels, without giving up what makes them themselves.

The Neighbourhood really has grown!

Useful Links

• Neighbourhoods Network

• Sid Sthalekar on LinkedIn

• Jill Burrows on LinkedIn

• Neighbourhoods LinkedIn

• Github

• YouTube

• Discord

• X

• Facebook

• Telegram

🌟 Latest News from hAppenings.community

From Anita - Several great experiences happened recently:

• Testing Unyt! Sam Turner, Sacha Pignot, and many others participated in the testing of how HOT swaps to Holofuel (and vice versa). What a fun time! As one of the least technical of the group, once I was helped (by Sam) to get all set up with my test “wallet”, it was all really slick. Super fast…omg! The Unyt team is doing an amazing job of being extremely thorough and testing rigorously. I was extremely impressed…and excited!!

• Testing Moss! I had the privilege of being an early tester with their latest release. They created a test group on Moss, and we enjoyed a Zoom call with Eric Harris-Braun to share wisdom about what our groups need for functionality on Moss. It’s an amazingly refreshing departure from the extractive thinking and siloing of other platforms. We’re going to “dogfood” (test) a hAppenings group on Moss!

• Testing our MVP, Requests & Offers! We had 13 folks in 2 days ask to join our test, which we’ve extended until May 3rd.

• Sam and I also had a great many Zoom meetings with amazing people with inspiring projects. We’ll morph some of those into articles down the road. Connecting with folks who feel Holochain can be a key piece of their stack is one of the greatest joys in my life!

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

🌟Latest News from the Holochain Ecosystem:

Article: Pre-Authorized: Claude Desktop and the Anatomy of Silent Consent by Mary Camacho from Cirdia

“The browsers Claude Desktop writes into are where women type the things they have not told anyone else.” Mary

That is the part too many companies still miss.

Women’s most private health writing no longer lives only inside apps.
It lives in portals. Browser tabs. Draft emails. Search bars. Messages never sent.

And when systems quietly create new paths into those spaces, this is no longer a question of convenience. It becomes a question of trust.

At Cirdia, Privacy-First Wellness starts with a simple belief: if a system touches Intimate Body Data, its architecture should make surprise impossible.

A forensic report on Claude Desktop's silent browser-bridge installation is the AI industry's Flo moment. The same consent-stretch pattern femtech already learned, now playing out at the assistant layer.

🌟Other Related News & Events

Request for Book Recommendations from All Tech is Human

All Tech Is Human curates an annual Responsible Tech Summer Reading List, and we’d love your help! Is there a book you would recommend? Key topic areas include:

• Responsible AI

• Trust & Safety

• Youth, Tech, and Wellbeing

• Public Interest Technology

• Tech & Democracy

• Tech Policy

Here's the link!

Article: Tuning AI to Think Like a Peacebuilder from Tech and Social Cohesion on Substack

Evaluation is revealing where AI falls short on Women, Peace and Security. Customization is showing how to fix it.

Article: 100+ Journalists Applaud the Internet Archive’s Role In Preserving the Public Record from FightfortheFuture.org

Over 100 journalists including Rachel Maddow, Cory Doctorow, and Ellen Nakashima have signed a letter to the Internet Archive celebrating the Wayback Machine as a crucial resource for their work.

Event: Introducing the DWeb Camp 2026 Venue: Alte Hölle, July 8-12, 2026

Alte Hölle is set on the grounds of a forest hotel dating back to the 1800s. Once a Stasi recreation site, Alte Hölle is now owned and inhabited by a collectivist community of friends who are turning this space into a sustainable venue for group events of technologists, hackers, artists, and activists.

Set on 100,000 square meters of lush meadows and ancient forests, just one hour southwest of Berlin, Alte Hölle is both a community and location with values that resonate deeply with DWeb, a place we are excited to nourish and leave richer than when we arrive.

Read the story by Tommaso Marmo, who interviewed some of the members of the community. Click here.

Early bird ticket sales, volunteer applications, and Call for Proposals are open on https://dwebcamp.org

Join hundreds of dreamers, builders, and hackers to envision the decentralized technologies we need today.⁠

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

🌟 Latest News from hAppenings.community

Update: Requests & Offers HC/MVP Alpha Test - 2nd Round!

We’re very excited to announce that Sam Turner and Sacha Pignot have worked their Holochain magic and we now are testing via Edge Node! All members are discoverable and we are proceeding with a second Alpha Test.

We’ll notify those of you who have already signed up to test, to see if you can help us with 20-30 minutes of your time! And, we’re inviting a few more folks to apply to test so we can have more feedback and group wisdom. Deadline to test is May 1, 2026.

If you’d like to apply to test, CLICK HERE- Thanks, everyone!!

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

🌟 Latest Official News and Events from Holochain, Holo & Unyt

Holochain Updates:

What does real-world battle-testing look like for an open-source distributed system?

It's Unyt running live transactions through Holochain at scale, reporting back what breaks, and us fixing it. Then running it again. Sometimes this happens multiple times a day.

When you're building infrastructure, theoretical tests tell you your code is correct on paper. Real-world tests tell you where your assumptions were wrong. You need both — the assurance of automated regression testing and the chaos of real people doing weird things that you never expected.

Unyt Accounting is trying to scale to thousands of users. Every time we hit a bottleneck in that test, we learn something. A memory leak we didn't see. A database query that scales badly. A network pattern that doesn't work at 2700 peers the way it works at 50. Each one of these is a data point.

This cycle (test, break, fix, test again) is what shapes development as we move from Q1 build-mode to Q2 delivery-mode. Not our roadmap in isolation. Not hypothetical scaling scenarios. The actual thing our partner is trying to do, and where it's failing.

That's not comfortable, but it's the only way to know if something actually works when it matters

The Holochain roadmap gets questions. This is an attempt to answer the common ones properly.

What is a point? A complexity unit. The team assigns 1–5 per task based on how complex the work is, not how long it will take. Software time estimates are notoriously unreliable, so the team doesn't make them. Instead, they track how many complexity points they complete over time and use that history to project forward. Current velocity: 34 points/week, measured as a 13-week rolling average.

Why does velocity change? Because it's empirical, not fixed. It can go up — a new team member joined recently, and careful use of AI coding tools has been contributing to a steady increase. It can go down — holiday seasons, unexpected complexity, anything that affects how much the team can take on. The 13-week average smooths out single-week variation and gives a more honest baseline.

Why does the Completed percentage sometimes go down? Because Holo and Unyt are actively alpha testing right now. Bug reports come in multiple times a day. When a new bug is filed, it gets added to the milestone scope — which can push the Completed count backwards. This looks discouraging on a dashboard. It's actually due diligence: the team is finding problems before production, not after.

Why doesn't Time Left decrease steadily? Three reasons.

First, these are teamwork weeks, not calendar weeks. The number answers: "If the team worked on nothing but this milestone, how long?" But the team is running multiple milestones simultaneously, with a different priority balance each sprint.

At the start of this year, the Holochain team moved from a flexible kanban workflow to structured, disciplined sprints. It's worth explaining what that actually means — because it changes how you should read their delivery timelines.

Here's how a sprint works: every two weeks, as the team approaches the end of a sprint, they estimate the remaining backlog. Tasks get assigned complexity points on a 1–5 scale — not time estimates. Time estimates for software are notoriously unreliable. Complexity is something developers can actually assess.

As we measure the number of complexity points we complete over time, we develop an understanding of our team's velocity that's based on data, not guesses. Tracking turns points into time — it lets our dev team understand what it means when it says "this task is three points". This is something we can use to make our time projections more reliable. The great thing is that this simple tool is resilient to team member variances, holidays, sick days, new hires, efficiency gains, show-stopper bugs, infrastructure downtime — all of these things get factored in automatically.

This is called evidence-based scheduling. it's used by some of the best dev teams, and we've used it ourselves for a good while now.

The magic happened when we adopted two-week sprints. Armed with the wealth of velocity data we'd already collected, we can now say "in this sprint, we have a decent degree of confidence that we can deliver x, y, and z." This makes our workweek more predictable, but it also helps projects building on Holochain, like Holo and Unyt, make more confident decisions based on our feature delivery timelines.

For anyone building on Holochain, or evaluating it as infrastructure, this matters. Predictable development cycles mean more reliable timelines for what you're building on top of. It doesn't mean Holochain ships on a perfect schedule — it means the team has real data to work from when they tell you when something is coming.

A footnote: Ironically, sprints might even be affecting our velocity for the better. We've seen our 13-week rolling average go up over the past couple months, from the mid-20s, up to 30, and now 34. Some of it might be due to a new team member or our careful use of AI tools to accelerate development and testing. But we have a hunch that the planned scoping has lowered our team's cognitive overhead, resulting in more clarity and focus.

Holo Hosting Updates:

Moss networks with Unyt billing automation are live. This solves a problem most infrastructure people understand only after it's too late: decentralized applications don't work with centralized cost structures.

Here's why: centralized intermediaries charge fees. Those fees compound through the stack. By the time users experience your application, the economics are so distorted by intermediaries that viability disappears. The cost structure kills the application before the architecture can.

Peer-to-peer billing changes that completely. Group networks now pay each other directly. Cryptographic accounting replaces invoicing. No intermediaries extracting fees at every layer.

This isn't just cost reduction. It's a fundamental shift in what becomes possible. Applications that required $100k+ in infrastructure budget to break even can now break even at $20k. Application categories that were economically impossible become viable. Business models that required venture capital become bootstrappable.

For builders and business leaders evaluating distributed applications, understand this: economics determines viability. It determines whether your application can scale sustainably. It determines your capital requirements and your path to profitability.

When distributed infrastructure eliminates intermediary friction, what new applications become possible? What business models become viable? What categories can you enter that centralized infrastructure made impossible?

This is why peer-to-peer economics matters strategically. It's not about saving a percentage on hosting costs. It's about unlocking entirely new categories of applications and business models that were impossible under centralized cost structures.

Holo Hosting + Moss + Unyt Accounting means the infrastructure for distributed applications is now economically viable. The stack is complete.

Unyt Accounting Updates:

The bridging mechanism into Unyt is not limited to blockchain tokens.

That's just the first implementation. Unyt is an accounting framework. It can reference any externally-defined asset, as long as that asset has a redemption guarantee satisfactory to the community. On-chain? Smart contract lock on the Ethereum side. Cryptographically enforced, verifiable, transparent. Off-chain? Traditional finance mechanisms. Escrow accounts. Surety bonds. Insurance policies. Auditable reserves you can inspect. The right approach depends on the asset and the trust requirements of your community. But the pattern is the same: bring assets in, let your community define how they move, guarantee redemption when needed. That's infrastructure for communities designing their own economics. The blockchain piece is the first entry point.

Join Holochain.org’s Developer Portal: https://developer.holochain.org/

For a complete listing of all of the Holo Ltd and Holochain Foundation events go to their ‘Upcoming Events’ page!

🌟Latest News from the Holochain Ecosystem:

Big announcement: Flowsta just launched Flowsta Sign It

With Sign It, anyone can easily sign any digital files to prove they were the creator.
The cryptographic signature is stored on a publicly accessible Holochain DHT. But also attach your terms, like:

1. How the work is licensed - Copyright, Creative Commons, public domain,etc..

2. Do you give permission for AI companies to train AI models on your work.

3. Has the work been generated by AI either in full, in part or not at all.

4. Are you available to be contacted regarding the work, either for commercial use inquiries or for any other reason you choose. Contact is through the Flowsta privacy-first contact system, so you never need to share your email publicly.

Anyone can verify it.

Drop the file at https://flowsta.com/sign-it/, or query the API directly — no key required, no rate-limit gotcha tier.

For AI training pipelines: hit the content-rights endpoint to check what the creator actually allows. We want this to become a standard. The schema is open.

Update: ValiChord — a scientific reproducibility verification system built on Holochain — ran its first fully automated end-to-end validation round using AI validators.

Three independent Claude AI agents each formed a reproducibility verdict on a computational study, without seeing each other's results. Both the researcher and the validators committed their results cryptographically before either side could see the other's. The outcome — Reproduced, ExactMatch, 3/3 validators — was written as a permanent, publicly readable record to a live distributed network.

No central authority. No authentication required to verify it. No way for either side to adjust their result after the other committed.

I believe this is the first time a blind commit-reveal reproducibility protocol has been run with AI validators on a live peer-to-peer network. I'm happy to be corrected.

🌟Other Related News & Events

Event: Introducing the DWeb Camp 2026 Venue: Alte Hölle, July 8-12, 2026

Alte Hölle is set on the grounds of a forest hotel dating back to the 1800s. Once a Stasi recreation site, Alte Hölle is now owned and inhabited by a collectivist community of friends who are turning this space into a sustainable venue for group events of technologists, hackers, artists, and activists.

Set on 100,000 square meters of lush meadows and ancient forests, just one hour southwest of Berlin, Alte Hölle is both a community and location with values that resonate deeply with DWeb, a place we are excited to nourish and leave richer than when we arrive.

Read the story by Tommaso Marmo, who interviewed some of the members of the community. Click here.

Early bird ticket sales, volunteer applications, and Call for Proposals are open on https://dwebcamp.org

Join hundreds of dreamers, builders, and hackers to envision the decentralized technologies we need today.⁠

Resource Referral from Lucas Tauil - Social Lean Canvas

From the Developer/Founder, Rowan Yeoman: “I’ve been developing the Social Lean Canvas as a set of free open-source tools for people building purpose-driven enterprises.

The aim is to support a new wave of social entrepreneurs and innovators working on problems they care about, and building financially sustainable ways to address them.

If you’re trying to turn an idea into a viable purpose-driven venture, or know someone who is, this might be useful: https://socialleancanvas.com/”

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

Have you ever wondered why some apps feel light and fast while others drain your battery and hog your memory? Or why some services respect your privacy while others seem designed to harvest everything they can? That weird experience of finding an advert for something you spoke to a friend about the other day in a totally different context and app? The answer often lives in decisions made long before you ever see a download button: the choice of tools a developer uses to build the software in the first place.

Sacha Pignot, lead developer of our Requests & Offers hApp, has published a detailed technical blueprint for building what he calls “sovereign software”. Entitled: The Compiled Stack: Building Sovereign Software with Rust, Tauri, Svelte, and Holochain, it’s a guide for developers, certainly, but the philosophy behind it matters for everyone who cares about how we interact with technology. This feels especially urgent as AI systems grow increasingly capable of correlating fragments of our digital lives across apps, devices, and contexts, assembling detailed portraits of us that we never consented to and can’t see.

His premise is simple: the tools you choose to build with shape what you can build. Choose tools designed around extraction and convenience, and you’ll produce software that extracts. Choose tools designed around efficiency, composability, and user sovereignty, and those values get baked into everything that follows.

A quick note on language: ‘sovereignty’ carries different connotations depending on where you’re reading from, from nation-states to fringe movements. In this context, it means something specific and practical: you control your own data, your own identity, and your own participation. No intermediary holds the keys. That’s the sovereignty Sacha is building toward.

His blueprint brings together a specific set of technologies, including Holochain, that share a common philosophy: compile away waste, make errors visible, respect the user. Many of them are already familiar to this community through projects like darksoil studio’s P2P Shipyard, Flowsta, and our own Requests & Offers. What Sacha has done is articulate why these tools belong together, and how the philosophy connecting them creates software that’s fundamentally different from what Big Tech produces.

We’re going to walk through the key ideas here in plain language. For those of you who are developers, or who want to use AI coding tools to explore building with Holochain, Sacha’s full article is a comprehensive technical reference. We’ll link to specific sections throughout so you can dive as deep as you’d like.

Ship Only What You Need

When you order something online and it arrives in a box three times its size, stuffed with plastic packaging, you know something’s wrong: the product works fine, but the delivery system is wasteful. Software works the same way.

Most modern apps ship with enormous amounts of code that you never need. The popular desktop framework Electron, used by apps like Slack and VS Code, bundles an entire web browser inside every application. That’s like delivering every parcel with its own postal van! It works, but your computer pays the price in memory, battery life, and storage.

Every tool Sacha has included in the stack takes the opposite approach: compile away what isn’t needed, and ship only the essentials. Rust, the programming language at the heart of the blueprint, produces lean, efficient code that runs close to your computer hardware. Svelte, the user interface framework, compiles the design into precise instructions before the app ever reaches you, delivering a fast and stable UI rather than interpreting it on the fly in that bundled browser. Even the styling system, UnoCSS, generates only the visual styles actually used, nothing unnecessary.

The result? Applications that start faster, use less memory, and respect the limited resources of the devices we actually own. For mobile users especially, this isn’t an abstract concern: it’s the difference between an app that drains your battery by lunchtime and one that hums along quietly in the background.

This matters beyond convenience. When software is lean, it can run on older devices and slower connections. It doesn’t demand the latest hardware to function. That’s a quiet form of digital inclusion: sovereign software that works for people, not just for people with the newest phones.

For the technical detail on Rust’s role across the stack, see Sacha’s section on [Rust as the Universal Backbone]. For how this extends to the browser through WebAssembly, see [WebAssembly: The Bridge Between Worlds].

One Language, Many Destinations

One of the persistent frustrations in software development is repetition. A team builds an app for desktop, then has to rebuild significant parts of it for mobile, then again for the web. Each version uses different tools, introduces different quirks, and requires separate maintenance. It’s like the difference between a 19th-century print shop, where every new format meant rearranging metal type letter by letter, and modern digital publishing, where the same manuscript flows to hardback, paperback, ebook, and audiobook from a single source. Some software development is still stuck in the print shop.

Sacha’s stack sidesteps this by anchoring everything in Rust. Because Rust can compile to native desktop code, to WebAssembly for browsers, and serves as the language Holochain is built on, a developer can write core logic once and deploy it everywhere. The rules that validate a transaction, the calculations that process your data, the security checks that protect your identity: written once, running identically whether you’re on a laptop, a phone, or a peer-to-peer network.

Tauri, the application shell in the stack, extends this further. Where Electron bloats with that unnecessary browser, Tauri uses your device’s own built-in web viewer and pairs it with a Rust backend. Think of a paperback that comes with an ebook link: same story, but you can read it on your phone on the train without needing to carry the physical book. Tauri delivers the same Rust app in whatever format suits the device, without the extra weight. With version 2, Tauri brings this approach to iOS and Android. This is the technology behind darksoil studio’s P2P Shipyard, which we’ve covered extensively, and it’s how Holochain apps are reaching mobile devices.

For users, the benefit is consistency. The app on your phone behaves the same as the one on your laptop because it genuinely is the same code underneath. For the ecosystem, it means small teams (like most Holochain dev teams are) can reach every platform without multiplying their workload.

For the full technical picture, see Sacha’s sections on [Tauri: The Shell That Gets Out of the Way] and [WebAssembly: The Bridge Between Worlds].

Built to Expect the Unexpected

Every book that gets published has been through an editorial process. Not just for spelling, but for consistency too: does chapter twelve contradict chapter three? Does a character who left the room suddenly speak from inside it? A good editor catches what the author can’t see, before the reader ever encounters it.

Sacha’s stack treats software errors with that same editorial discipline. Most applications handle errors as afterthoughts, bolted on where problems have already occurred rather than designed in from the start. Something goes wrong, the app crashes, and you’re left staring at a cryptic message or, worse, no message at all.

In Rust, the language itself refuses to compile code that ignores potential failures. If a function might fail, the developer must explicitly account for that possibility before the code will even run. It’s an editor who won’t send the manuscript to print until every loose thread is resolved.

This philosophy carries through to the frontend. TypeScript, the language used for user interfaces, is paired with a library called Effect that brings the same discipline to the browser. Every operation that could fail, and in a distributed system that’s most of them, declares exactly how it could fail and what should happen when it does. A peer goes offline? That’s a specific, named scenario with a planned response, not a generic crash.

For users, this translates to apps that degrade gracefully rather than fail catastrophically. Instead of freezing or losing your work, a well-built application in this stack can tell you precisely what happened and, in many cases, recover on its own. Peer-to-peer connections behave differently from centralised servers: rather than one point of failure that takes everyone down at once, the network flexes around individual nodes going offline. Rust’s definitive error handling is built into the stack at every level, making dynamic flexibility seamless. This kind of resilience isn’t a luxury. It’s essential.

It is one of the less visible but most consequential choices in the stack. As a user you may never notice it working, which is exactly the point.

For Sacha’s detailed comparison of error handling in Rust and TypeScript, see [Functional Programming: The Unifying Methodology].

Your Data, Your Device

Sacha’s article frames the Holochain stack in a way that’s worth highlighting: not as a monolithic standalone technology but as a natural destination for a stack built around sovereignty.

He has explored this idea further in his blog calling it ‘fractal sovereignty’, a framework where local autonomy and global cooperation aren’t opposites but nested layers of the same system. It’s a useful lens for understanding why Holochain’s architecture works the way it does.

Every principle we’ve walked through so far converges here. Holochain is written in Rust, so it inherits that efficiency and cross-platform reach. Holochain applications compile to WebAssembly, so their logic can run in browsers. And Holochain’s distributed architecture means your data lives on your device, validated by peers on the network, rather than sitting on a corporate server waiting to be monetised or breached.

To return to our publishing thread: most of today’s software is like a library where you can borrow books, for an ongoing fee, but never own them. The publisher decides what’s available, can withdraw titles without notice, and keeps detailed records of everything you’ve read. Holochain is more like owning your own collection outright, stored in your home, with the ability to lend to trusted friends directly. No intermediary deciding what you can access or tracking what you do. It’s worth noting: this sovereignty holds as long as the tools integrated with the hApps respect the same principles. An AI assistant that sends your data to commercial servers would undermine it, which is why locally run and privacy-preserving AI solutions are an active area of development in the ecosystem.

What Sacha articulates clearly is how Holochain’s internal architecture mirrors the modularity of the wider stack. Holochain is itself built in layers: connection protocols at the base, data synchronisation through Kitsune (the gossip layer that spreads information between peers), validated personal ledgers running sandboxed applications, and user interfaces on top. Developers can swap components at any layer for their specific use case. Some have replaced the connection layer with alternative networking tools. Others have pulled Kitsune out of Holochain entirely to power their own projects, like taking a chapter from one book and building an entirely new story around it.

This flexibility is why asking “what works with Holochain?” produces such varied answers. Web developers can scaffold familiar interfaces using React or Vue. Systems architects can connect to it as a library. As Eric Harris-Braun, Holochain’s co-founder, has noted: ‘you can use different technologies at any place in the stack. Holochain meets developers wherever they are, with whatever tools they already use.’

For non-developers, the takeaway is this: the apps being built on Holochain aren’t fragile experiments tethered to a single framework. They’re being constructed with the same modular, composable principles found in foundational tools like Linux and Git, software that proved robust, world-scale infrastructure can be built without extracting from users. Holochain’s stack carries that same philosophy into the applications you’ll use every day.

For the full technical detail, see Sacha’s section on [Holochain: Distributed by Design].

Familiar Foundations

If this philosophy sounds familiar, that’s because you’ve already seen it in action across these pages. Darksoil studio’s P2P Shipyard uses Tauri and Rust to get Holochain apps onto every device. Flowsta uses Holochain’s cryptographic architecture to give you single sign-on without surveillance. Our own Requests & Offers hApp, which Sacha is building, uses this very stack: Rust and Holochain on the backend, Svelte and TypeScript with Effect on the frontend. These aren’t isolated choices. They’re expressions of a shared architectural philosophy that Sacha has now documented as a coherent whole.

A Blueprint You Can Use

Sacha’s article isn’t just a manifesto. It’s a practical reference. If you’re a developer exploring Holochain, it’s a detailed guide to building your first app with a principled, production-ready stack. If you’re not a developer but you’re curious about building tools for your community, this is where it gets interesting: AI coding assistants can work with detailed architectural blueprints like Sacha’s to help you scaffold applications. Hand his article to an AI tool, describe what you want to build, and you have a foundation to start from. It won’t write your app for you, but it dramatically lowers the barrier to experimenting with sovereign software.

Whether you’re here to understand what the Holochain ecosystem is or ready to start building, Sacha’s article is worth a deep dive to appreciate the elegant efficiency baked into the stack.

Read the complete technical blueprint: The Compiled Stack: Building Sovereign Software with Rust, Tauri, Svelte, and Holochain

Useful Links

• Sacha’s Full Article

• Sacha’s Alternef Digital Garden

• Requests & Offers GitHub

• darksoil studio/ p2p Shipyard

• Flowsta

• Holochain Developer Portal

🌟 Latest News from hAppenings.community

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

🌟 Latest Official News and Events from Holochain, Holo & Unyt

Holochain Updates:

Most platforms promise security. Holochain proves it mathematically.

There's a meaningful difference. Promises can be broken. Code can be compromised. Architectures can be exploited. But mathematical proofs persist. They don't depend on vendor credibility or marketing claims. They depend on cryptography.

Holochain's 0.6 release hardens the cryptographic validation layer through warranting. Agent-centric architecture now enforces security not through claims, but through mathematics. Validation rules are cryptographically verified. Invalid data triggers network-wide immune response. Not because we say so. Because the cryptography forces it.

This matters for institutional adoption. Institutions don't trust marketing claims. They trust mathematical proof. When security is enforced by cryptography rather than promises, risk decreases dramatically. Your infrastructure security doesn't depend on vendor reputation. It depends on mathematics that can be audited, tested, and verified.

For builders and decision-makers evaluating distributed infrastructure, this distinction is critical. Production-ready infrastructure proves security through mathematics, not claims. Experimental infrastructure makes promises and hopes the architecture holds.

Holochain crossed that threshold. The cryptographic validation layer proves the architecture works as intended. That's the foundation for institutional trust.

This is what infrastructure maturity looks like: when security moves from theoretical to cryptographically enforced, the infrastructure becomes credible for serious applications. That credibility unlocks institutional adoption. That adoption drives ecosystem growth.

Holochain reached production credibility because mathematical proof replaced marketing claims.

Bootstrap service ownership changes something structural about network integrity.

Here's the problem most distributed systems face: you can only control access at the application layer. The network itself is open. Bad actors join, spam the DHT, harvest addresses, waste computational resources. You build walls at the top. The foundation is still porous.

What if the bootstrap server, the actual entry point, belonged to the hApp instead?

Each application would control enrollment at the protocol level. When an agent installs an hApp, they're joining a network whose gatekeeper is the hApp itself. You can't join the DHT without the bootstrap server. You can't get past the bootstrap server without the hApp's approval. No application-layer policy can work around it.

This combines with warranting, which we talked about before. Every agent has a source chain, a cryptographically signed record. When an agent claims membership, peers validate that claim against the app's rules. Invalid claims aren't just rejected. They're recorded. The network-wide immune response kicks in.

The tradeoff: this means each hApp maintains its own bootstrap infrastructure. That's work. It's operational cost. But for any application where integrity matters at enrollment, it's cheaper than dealing with spam, resource exhaustion, or reputational harm.

This is what happens when access control is architectural. It doesn't need policy at the top because the structure itself is selective.

Most infrastructure projects fail slowly.

Not through one catastrophic mistake, but through accumulated inconsistency. Promises that slip. Timelines that stretch. Teams that burn hot for a quarter and then go quiet.

It's hard to spot until the pattern is obvious, and by then, trust is already spent.

We track velocity differently. Not by what we shipped last sprint, but by where our 90-day rolling average lands. Right now it's holding at 31 points. Individual weeks swing; we've had weeks at 16 and weeks at 60. The average is what tells the real story.

HC 0.6.1 has four weeks of teamwork remaining. Not four calendar weeks, engineering work doesn't map cleanly onto calendars, and anyone who's shipped real software knows that. What we can say is that the work is scoped, the team is moving, and the velocity is consistent.

For builders evaluating a framework to stake real applications on, that consistency is the signal worth watching

Holo Hosting Updates:

Infrastructure traction through community adoption

The Holo EdgeNode Support channel saw 30% growth in a single day. This is not a marketing metric. It is a traction signal.

When developers and operators join a support channel unprompted, it means the infrastructure solves a problem they actually have. EdgeNodes lower the friction for hosting hApps. The community is responding by adopting, testing, and building on top of it.

This is how you identify infrastructure with real product-market fit: builders show up, ask hard questions, and solve problems together. No hype required.

Most infrastructure projects fail slowly.

Not through one catastrophic mistake, but through accumulated inconsistency. Promises that slip. Timelines that stretch. Teams that burn hot for a quarter and then go quiet.

It's hard to spot until the pattern is obvious, and by then, trust is already spent.

We track velocity differently. Not by what we shipped last sprint, but by where our 90-day rolling average lands. Right now it's holding at 31 points. Individual weeks swing; we've had weeks at 16 and weeks at 60. The average is what tells the real story.

HC 0.6.1 has four weeks of teamwork remaining. Not four calendar weeks, engineering work doesn't map cleanly onto calendars, and anyone who's shipped real software knows that. What we can say is that the work is scoped, the team is moving, and the velocity is consistent.

For builders evaluating a framework to stake real applications on, that consistency is the signal worth watching

There's a principle in infrastructure that most people miss: who controls the infrastructure controls the economics of the applications running on it. This determines everything about your business model, your independence, and your long-term viability.

Holo Hosting Edge Node is production-ready. What that means: operators can deploy hosting nodes on any OCI-compliant infrastructure. HoloPorts. Custom hardware. VPS. Whatever you own or control.

This seems like a technical feature. It's actually a business model decision.

When your infrastructure depends on a vendor-controlled platform, your economics depend on their pricing, their policies, their roadmap. You're one API change away from restructuring your costs or your architecture. Your competitive advantage becomes hostage to platform decisions you don't control.

When your infrastructure is yours—deployed on hardware you choose, running software you control—your economics and your competitive advantages stay yours. Your margins aren't at risk from platform policy changes. Your feature roadmap isn't constrained by vendor limitations.

For CTOs and infrastructure architects building distributed applications, this distinction matters enormously. Edge Node production-readiness means distributed apps can now achieve real infrastructure autonomy. You own the stack. You control the economics. You determine the competitive boundaries.

This is why Holo Hosting matters beyond hosting. It's about who owns the infrastructure that runs distributed applications. When operators own the infrastructure, distributed applications achieve real independence. That independence drives real viability.

The crowd is the cloud. That's more than a tagline. It describes who captures value, who controls decisions, and who owns the future of applications built on the platform.

Join Holochain.org’s Developer Portal: https://developer.holochain.org/

For a complete listing of all of the Holo Ltd and Holochain Foundation events go to their ‘Upcoming Events’ page!

🌟Latest News from the Holochain Ecosystem:

Update from Moss.Social

The chat history your team built inside Discord belongs to Discord.

Not legally, practically. It lives on their servers. You can request an export, but there's rarely somewhere meaningful to take it. The context, the decisions, the institutional memory: all of it stays tethered to that platform's infrastructure.

Moss Vines v1.26 shipped this week with Discord chat import. Users can now export their Discord conversation history and bring it into Moss Vines, a group chat tool built on Holochain.

Moss Vines is an ecosystem project, not a Holo product. But it runs on the same agent-centric architecture that Holochain apps use, which means your data is stored relative to your identity, not on a central server controlled by the service provider.

The harder problem with data portability has never been the concept. It's building a stack where portability is structurally possible at all. Distributed hosting, community-operated nodes, and data that doesn't depend on a central authority to exist.

For organizations thinking about infrastructure resilience, this is worth paying attention to. When the platform your community lives on changes its terms or pricing or shuts down, what happens to the history you built there? That question used to have one answer.

Blog: Nondominium - A Coordination Layer for Parallel Infrastructure by Tiberius Brastaviceanu of Sensorica

Financializing physical assets is a bad idea. Here's how we approach RWA (Risk-Weighted Assets).

Update: Milestone for Valichord from Ceri John

Excited to share a milestone for ValiChord — our scientific reproducibility verification system has just had its first external integration merged into Feynman.
Feynman, created by Advait Paliwal, is an open-source AI agent platform that autonomously replicates scientific studies, audits research code, and conducts deep literature research. It's a serious tool gaining real traction in the research community.

With this integration, Feynman now extends into verification — AI agents can act as autonomous reproducibility validators, sealing blind commitments, waiting for phase transitions, revealing findings, and generating tamper-proof verification records. The two systems cover the full pipeline: Feynman handles discovery and replication, ValiChord handles cryptographic verification of the result.

🌟Other Related News & Events

Article: Technology as if Life Mattered — Part I
Civilizational Transitions Require New Systems by Steve Melville

What becomes possible when we stop denying collapse—
and start building for what comes after?

That’s the starting point for: Technology as if Life Mattered — Part I
Civilizational Transitions Require New Systems

I recently published the first installment of this five-part series exploring the technological substrate needed for a civilization aligned with life.

If you’re working on questions of regeneration, systems change, the future of coordination, or the technological foundations that support them:
I’d be interested in how this lands.

Documentary: “The AI Doc” with Tristan Harris and Asa Raskin from the Center for Humane Technology

"The AI Doc" premiered on Friday(3/27) and audiences across the country came out to see this critical film. If you haven't seen it, there's still time.

The film clearly lays out the stakes and the urgency of the challenge ahead of us. Right now, a handful of companies are deciding how AI will shape humanity's future. They're accelerating but no one's at the steering wheel. This is our chance to take the wheel and steer toward a better future.

Seeing the film is the first step towards that better future. But now comes the hard part — taking action. That's why we've released "The AI Roadmap," a report outlining seven principles for how AI should be built, deployed, and governed. You can read the report on our website.

Together, we can shift the direction of AI.

Event: ATIHx Oakland

On April 30th, the All Tech Is Human community is coming together for an ATIHx gathering--our first to happen in Oakland. Join in!

ATIHx Oakland is being led by Aram Fischer. ATIH's Sherine Kazim will be there, along with dozens of people from our large Slack community (anyone can join).

ATIHx gatherings are independently-organized meet-ups that leverage our organization's connectivity as the global hub for the Responsible Tech ecosystem. Recent ATIHx gatherings have happened in Barcelona, Toronto, Philadelphia, and more!

We're thrilled that the Oakland community is coming together. Feel free to bring others who have not previously been involved with All Tech Is Human but are looking to be involved with the Responsible Tech movement.

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

The Missed Note

Most published biomedical computational research has never been independently verified. You read that right, and it’s not because verification is impossible but because there’s been no validation infrastructure to make it happen.

The numbers are stark. Between 75% and 90% of published biomedical studies cannot be reproduced by independent researchers. The annual cost of that failure runs to an estimated $28 billion in the US alone, with hundreds more spent globally, that’s poured into findings that don’t replicate, treatments that don’t work, and policies built on false foundations.

Ceri John, a brass music teacher and award-winning documentary filmmaker from Burry Port in south Wales, is not a developer or scientist. He is someone who sees patterns, and he saw a structural gap that the scientific establishment had failed to close.

“A chord requires multiple notes sounding together,” he explains. “Agreement is harmony. Disagreement is dissonance. Both carry information.”

That musical instinct has become ValiChord: a distributed validation infrastructure for computational biomedical research, built on Holochain, currently targeting UKRI’s Metascience funding programme. It may be the most quietly ambitious project in the ecosystem right now.

Variations on a Theme

This isn’t a new problem, and it isn’t for lack of trying. The internet and data repositories were supposed to make research materials accessible, but 70% of researchers have failed to reproduce another scientist’s experiments, leaving validators with nothing to work from. Blockchain-based verification promised a solution until it collided with GDPR and volatile gas fees: patient data cannot be deleted from an immutable ledger, making blockchain architecturally incompatible for the sensitive data used in medical research. Journal reproducibility badges gesture toward openness without requiring actual verification: only one in three reviewers even checks whether a final paper matched its registered protocol. Each solution addressed a symptom: none addressed the systemic infrastructure gap.

Ceri saw that validation is valued labour, not invisible service. By creating incentives for validators to be compensated, credited by name, and recognised through formal attestations, the system becomes anti-fragile. He has designed ValiChord so that honesty is simply the most rational strategy, not because validators are assumed to be virtuous, but because the architecture makes gaming the system unrewarding.

Even the most significant advances in research integrity struggle with this. Registered Reports, a peer review format now used by over 300 leading journals worldwide and developed by The Center for Open Science, was specifically designed to eliminate the incentive to p-hack, cherry-pick, or bury inconvenient findings. Yet even here, only one in three reviewers checks whether a final paper actually followed its registered protocol (Registered Reports FAQs #1 & #14). The intent can be verified. The execution is another matter entirely. That is the gap ValiChord was built to close.

The Orchestra

Think of ValiChord as an orchestra. Each section plays its own part independently, contributing its distinct sound to the whole. When performing together the sections create harmony, and when you hear something jarring, that’s not a flaw, that’s music expressing something important. Dissonance is information.

ValiChord’s architecture has four sections.

• Researcher — The String section carries the melodic foundation: the researcher holds their own private score, raw data, analysis protocol, pre-registered hypotheses, sealed cryptographically on their own device, never shared directly.

• Validators — The Brass section are the validators, powerful and independent, each working in their own private computational workspace, reproducing the analysis without seeing what their peers are finding.

• Attestation — The Woodwind section is the attestation layer, the connective tissue where validators publish their cryptographic commitments simultaneously, voices coming together for the first time.

• Governance — Conductors are the governance layer, the organisational backbone that assembles what the orchestra produces into a permanent public Harmony Record, accessible to journals, funders, and institutions worldwide.

ValiChord’s Harmony Records aren’t binary pass/fail verdicts or pop charts out of 10, they’re full of analytical feedback: who agreed, who didn’t, and most crucially, precisely where the divergence lies. Harmony means the research progresses: to the next funding stage, to wet lab testing, to publication with verified foundations.

Disharmony is equally valuable, returning specific, granular, validated feedback on exactly where the analysis needs work before it cascades through the medical pipeline, from lab bench to patient bedside.

The Pit

In a theatre, the orchestra pit sits below the stage, invisible to most of the audience. The performance above depends entirely on what happens below. Without the pit, there is no show.

Holochain is ValiChord’s pit.

ValiChord is built on Holochain, a cryptographically validated peer-to-peer distributed data and application framework. Each participant, whether researcher, validator, or institution, holds their own data locally on their own device. What travels across the network is not the data itself but the cryptographic proof that it exists and hasn’t been tampered with. The voice, not the score. There is no central server holding everyone’s information, no single filing cabinet that can be breached, subpoenaed, or shut down.

Similar to how no one member of the orchestra holds the full score, nor is any one musician required to hold every part in the set, Holochain distributes encrypted proofs across the computers, or nodes, of its participants in a network structure called a Distributed Hash Table (DHT). Each node holds only what its participant needs: private data stays private, and only the cryptographic proofs of validation travel across the network, recorded in each participant’s own locally stored immutable ledger. Just as each musician has their own part and together they combine to produce the piece, each node contributes its part to the whole. The complete validation data exists in no single hand, only as an ensemble, publicly shared as a live performance or a recording of it, known as the ValiChord Harmony Record.

For biomedical research, this matters enormously. Patient data, clinical trial results, and sensitive research materials are subject to GDPR and a web of international privacy law. Blockchain, despite its promise, failed this test. Its defining feature, an immutable global ledger, means data written to it cannot be deleted, and it requires some form of global single source of truth which replicates data in nodes outside of user’s control. That makes it architecturally incompatible with medical research. Holochain resolves this not through policy or legal workarounds but through architecture. Sensitive data never enters the shared network because the system was never designed to put it there.

Going for a Song

Because each participant maintains their own local ledger rather than competing to write to a single global one, there is no mining, no global consensus mechanism, and no gas fees. Validators can self-host or cooperate headlessly with the encrypted data, or be compensated for micro-transactions at fractions of a penny, making the economics of distributed scientific verification viable for the first time.

The Conductor

In a striking symmetry of language, Holochain’s own framework uses the term ‘conductor‘ to describe the process that coordinates its distributed performance. It is not a metaphor borrowed from music. It is the actual name of the technical process that manages the running of the application’s code.

That Holochain’s founders chose that word, and that Ceri, a musician, independently arrived at the same conceptual territory when designing a system for scientific integrity, is either a remarkable coincidence or evidence that elegant systems, whether musical or computational, tend toward the same language.

Recently, Ceri and Art Brock, Holochain’s co-founder, connected. Ceri shared his architecture. Art listened and brought the deep expertise of someone who has spent more than a decade thinking about how distributed systems can serve human coordination. The conversation that followed made ValiChord stronger.

“When Art looked at my documents, he was kind but honest. Everything you’ve built here is still thinking in centralised terms, he told me. That’s not how Holochain works, and more importantly, it’s not how this problem works.

Then he asked one question that stopped me in my tracks: where are your membranes?

He explained that in Holochain, each part of the system isn’t a layer sitting on top of another — it’s a separate enclosed world, with its own participants, its own data, its own rules about what can enter and leave. The key insight was that sensitive data never crosses those boundaries at all. The only thing that travels between worlds is a cryptographic fingerprint. A proof that something exists, not the thing itself.

It meant rebuilding. But what I ended up with was genuinely more secure, more private, and more honest than what I’d started with.

Before that conversation, ValiChord was a good idea wearing Holochain’s clothes. After it, Holochain’s way of thinking about the world had become the idea itself.”

Rondo

In classical music, a rondo is a form where the main theme keeps returning, each time enriched by what came between. ValiChord’s economic design works the same way.

As Harmony Records accumulate and funders gain confidence in ValiChord’s verified science, the incentive for validators to participate grows. More validators means more studies verified, more credible investment opportunities, and more trustworthy data for everyone who depends on it. Each turn of the cycle strengthens the next. This is the virtuous cycle at the heart of ValiChord’s design; a self-sustaining loop where credibility generates participation, participation generates credibility, and science gets better in the process.

And for funders? ValiChord replaces the unicorn model, pouring money into a broad base of speculative projects in the hope of a breakthrough, with the confidence to strategically invest based on real-time, cryptographically verified research they know can be replicated at each stage. ValiChord delivers a fundamental prerequisite for the scientific method.

Building Towards the Crescendo

ValiChord is not just a conceptual whitepaper. It’s a validated, production-quality architecture with 669 commits, nearly 28,000 lines of code across Rust, TypeScript, and Python, and 73 integration tests passing across all four DNAs. The hard thinking has been done. The infrastructure is built and tested.

The architecture has been reviewed by Art Brock, Paul D’Aoust of the Holochain Foundation, and Joel Marcey of the Rust Foundation. Three independent technical experts, each bringing deep domain knowledge, each confirming the foundations are sound.

What comes next is Phase 0: a 12-month empirical study that answers the one question that can only be answered in practice. How long does it actually take to verify that published research holds up? Phase 0 finds out.

Eight to ten computational studies. Sixteen to twenty validation events. Qualified researchers reproducing published analyses, recording their time across four categories: environment setup, data acquisition, code execution, and troubleshooting. The results feed the difficulty prediction model that makes ValiChord’s compensation tiers credible and its economics sustainable.

Ceri is actively targeting UKRI’s Metascience grant programme to fund Phase 0. Beyond testing and hardening the infrastructure Phase 0 will also scope the necessary validators specific qualification requirements, compensation tiers, and difficulty assessments the economic model depends on. These are questions only real-world evidence can answer.

ValiChord is at the threshold. The score is written. The orchestra is assembled. The house lights are dimming. Here at hAppenings we’re excitedly clapping in anticipation as the orchestra warms up!

Play Your Part

You don’t need to wait for Phase 0 to engage with ValiChord. Right now, today, Ceri has built two tools that any researcher can use immediately.

Step 1: ValiChord at Home is a free, private reproducibility readiness check for your code repository. Twenty-four questions across six categories: Documentation, Dependencies, Environment, Data Access, Code Organisation, and Self-Verification. Each is designed to show you exactly what a validator would need to find in your repository. It produces a diagnostic report, not a score. It runs entirely in your browser. Nothing is tracked, stored, or shared.

Step 2: Repository Readiness Check. Upload your repository as a zip file and ValiChord analyses it against 126 reproducibility failure modes across documentation, dependencies, environment setup, data provenance, and more. It returns a full diagnostic report with findings categorised as Critical, Significant, or Low Confidence, plus starter drafts for any missing documents. It’s a static analysis tool that tells you precisely where the gaps are before a validator finds them.

• If you’re a researcher, start with Step 1 today.

• If you’re a validator, developer, or institution interested in contributing to Phase0, Ceri wants to hear from you.

• If you’re a funder who recognises what verified science could mean for your portfolio, the door is open.

The orchestra is assembled. Your part is waiting.

Useful Links

ValiChord Website
ValiChord at Home
Valichord Github
Ceri Johns LinkedIn
Registered Reports
Center for Open Science

🌟 Latest News from hAppenings.community

If you’re new to “What’s hAppening” and the hAppenings Community, WELCOME! We invite you to explore our website, hAppenings.community, our Holochain Project Database (searchable), and we would love if you would join our Discord Server!

Also, if you’ve you’d like to see a type of retrospective of Holochain projects, check out the back issues of this newsletter. It’s rather amazing! And we always appreciate new and paid subscribers (who help keep the lights on!) Thanks, Anita Halcyon & Sam Turner!

Subscribe now

🌟 Latest Official News and Events from Holochain, Holo & Unyt

Blog: Who Gets to Join? Introducing Configurable Membranes in Unyt by Matthew Schutte of Unyt Accounting

Unyt now supports configurable membranes, a joining workflow that gives network operators control over who participates and under what conditions.

The workflow runs on checkpoints: modular challenges a prospective participant must complete before joining. Use one, stack several, or configure none and let everyone through.

Available checkpoint types:
- Fully open — no authentication required
- Invite-gated — single-use or shared invite code
- Email-verified — confirms the user controls the address
- Agent whitelisting — pre-approved cryptographic keys, for enterprise tools, IoT, or machine-to-machine interactions
- Operator approval and KYC — human review or external KYC provider, for regulated applications

For more, click here.

Unyt Updates:

What does it mean to build your own currency?

Most people think of currency as something governments or banks create. But currency is just a shared system for tracking value, and there's no reason communities can't build their own.

Circulo, built on Unyt, lets any group do exactly that. A neighborhood, a cooperative, a network of DePIN devices, each can define units of account, set the rules, automate the billing, and trade across other currency systems without asking permission from anyone.

Smart Agreements handle the logic: configurable templates, community-specific rules, verifiable execution records. Business users configure them and reconfigure them as they learn more. No code required.

This is what economic infrastructure looks like when it's built for participants instead of intermediaries.

Real life doesn't happen in "Trading Pairs."

The current Web3 and TradFi systems force us into bottlenecks. If you want to exchange assets, you usually have to convert them into a base currency (like USD or ETH) first.

But what if a logistics company wants to settle a debt by paying partially in fuel credits, partially in labor hours, and partially in cash?

Unyt handles this natively. Our engine allows for complex, bundled trades between peers. It either all succeeds at once, or it all fails. No intermediate conversions, no massive liquidity pools needed. Just pure, frictionless commerce.

Holochain Updates:

Ecosystem Momentum Update

Things are moving this week:

The development roadmap is tracking at 31 points per week. Steady velocity toward release in about 4 weeks. Wind Tunnel testing infrastructure is maturing with better data visualization and expanded test coverage. Unyt shipped pricing oracles for hApps that need external market data.

We also published a blog post exploring how agent-centric architecture and AI agents can work together — two different meanings of “agent” that might converge in interesting ways.

This is what progress looks like when you focus on the work instead of the noise. Consistent, transparent, moving forward.

See Roadmap Here

Engineering Discipline: How Testing Infrastructure Matures

Holochain’s Wind Tunnel testing infrastructure is approaching the release milestone with steady progress on what might be called unglamorous work — but it’s the work that matters.

Recent focus areas:

Data visualization: Polishing the display layer for performance reports and summary data. When you’re testing distributed systems, understanding the results clearly is essential for making informed decisions.

Test coverage expansion: Adding comprehensive tests for fundamental operations like signals. Building reliability means systematically verifying behavior across the entire stack.

Real-world patterns: Unyt patterns are receiving significant attention in the test suite — verifying that actual hApp use cases perform reliably under production-like conditions.

The roadmap continues tracking at 31 points per week. This is the engineering discipline that separates experimental frameworks from production-ready infrastructure.

Solid testing practices may not be exciting, but they’re what make the difference when real applications depend on your platform.

We’ve been conditioned to think software only matters if it scales to a billion users on a single database.

But true resilience starts small. We don’t need to “save the world” with a monolithic platform. We need to save our own communities with tools we actually control.
Holochain gives you the framework to build at the human scale. Your rules. Your infrastructure.

Holo Hosting Updates:

There's a growing gap between the internet people want and the internet that currently exists.

More people are asking: who owns this infrastructure? Who controls the data? What happens when the platform decides to change terms, shut down, or monetise the community it built?

These aren't abstract questions anymore. They're prompting real decisions about which tools to trust and whether it's worth the effort to run something yourself.

The Holo ecosystem is built around a straightforward premise: distributed infrastructure works better when more people participate in it. Not as users. As operators.

That shift, from consuming infrastructure to running it, is what community-owned networks are actually about. And the tooling to make that accessible is more mature than it's ever been.

The Holo Node ISO, Edge Node support for ARM hardware, and in-browser node management are all steps in the same direction: lower the barrier until participation is a realistic choice for far more people.

Network resilience isn't a feature you add at the end. It's something you test under real conditions, continuously.

Holochain's Wind Tunnel is a network stress-testing environment designed to do exactly that, and anyone can run a node in it, contributing real data on how peer-to-peer systems hold up under load.

There's an underappreciated model here: distributed testing by a distributed community. The people running the network are also the people stress-testing it. That alignment is by design, not coincidence.

If you're thinking about what genuinely resilient distributed infrastructure looks like, and what it takes to build it, the Wind Tunnel approach is worth understanding.

Read More Here
Watch a Video Here

Something genuinely exciting is happening in our community right now.

Members are figuring out how to run Holo Edge Nodes on Android devices, and they’re documenting everything as they go. Tutorials, tips, workarounds, and troubleshooting. The knowledge base is growing quickly, built entirely by people sharing what they’ve learned.

There’s a dedicated Telegram group where all of this is happening:

Both x86 and ARM architecture are supported, so a wide range of hardware can run a node, including Android devices.

If you have an Android device and want to put it to work on something meaningful, this community is the place to start. And if you’re just curious to see what people are building, it’s a great follow.

Running a Holo Node just got a lot more accessible.

The new Holo Node ISO comes with an in-browser Node Manager, meaning you set it up and manage everything through your browser, no technical command-line knowledge needed.

It also supports a personal AI assistant of your choosing, running on your own hardware. And it works with both x86 and ARM hardware, including a lot of the affordable mini-PCs and small computers that have become popular in recent years.

If you've been curious about running your own node but assumed it was too technical or too complicated, this update was made with you in mind.

Join Holochain.org’s Developer Portal: https://developer.holochain.org/

For a complete listing of all of the Holo Ltd and Holochain Foundation events go to their ‘Upcoming Events’ page!

🌟Latest News from the Holochain Ecosystem:

Update: AD4M v0.12.0-rc2: AI Agent Integration as First-Class Infrastructure

We’ve (Coasys) just released the second release candidate for AD4M v0.12.0. While technically labeled an RC, the scope of changes between RC1 and RC2 reflects the rapid pace of development as we move toward production readiness.

Key technical achievements:
OpenClaw Plugin Integration
AI agents can now connect to AD4M with a single command. The plugin handles executor lifecycle, authentication, and MCP tool exposure — enabling AI agents to participate in decentralized social coordination graphs with zero configuration overhead.

Model Context Protocol Server
Our embedded MCP server introduces a novel capability: register a SHACL data model and receive auto-generated CRUD tooling. This abstracts AD4M’s complexity for AI agents while preserving the full expressiveness of the semantic graph layer.

Rust Architecture Migration
We’ve completed the removal of the TypeScript executor (4,600 lines deleted), porting the LanguageController to pure Rust. This delivers proper language sandboxing, reduced memory footprint, and simplified deployment — critical for production hosting environments.

Hosted AD4M Foundations
Credit systems, compute tracking, and administrative interfaces are now in place, enabling AD4M-as-a-service for teams and organizations.

Additional capabilities are in active development. This release represents infrastructure for the next phase of decentralized collective intelligence.

🌟Other Related News & Events

Event: Tickets DWeb Camp 2026 are up for sale! July 8-12, 2026 at Alte Hölle, Germany

DWebCamp 2026: Root Systems is a five-day gathering in nature for builders and dreamers dedicated to building a decentralized web (#DWeb).

Join us at Alte Hölle, Germany, on 8–12 July 2026 to shape the decentralized, democratic, and open #technology we truly need.

• Get your Early Bird ticket

• Submit your idea to the Call for Proposals

• Be the backbone of our camp and apply to volunteer!

Announcement: Hylo.com receives the 2026 Pizzigati Prize

Hylo is incredibly honored and humbled to receive the 2026 Pizzigati Prize, given out by NTEN each year at its annual Nonprofit Technology Conference!
The prize celebrates software developers who create open source apps and tools that nonprofit and advocacy groups can put to good use.

"The Pizzigati Prize for Software in the Public Interest honors the life and legacy of Antonio “Tony” Pizzigati, an early advocate of open source computing. Tony never had a chance to fulfill his computing dreams, so the prize was created to help others realize theirs. The Pizzigati Prize celebrates open-source projects that power nonprofit impact."

We are committed to fulfilling the the legacy of Tony Pizzigati and continuing to turn Hylo into a powerful and flexible tool for non-profits and purpose driven organizations and communities of all kinds to effectively collaborate, coordinate, and bring their missions to life.

Find more information and the announcement video here

Handout: Pause. Before You Build: 10 Mindsets for Wise Innovation in the Age of AI by Andrew Murray Dunn

Because when AI can make anything, the hard problems shift from technical and operational to cultural and relational.

The best time to engage these ideas is at the very beginning. The second best time is now.

This handout is just a snapshot of a 24 page booklet that we worked with during the workshop (see link in comments), which is a condensed and localized version a ~100 page book I’m working on around right relation to vision.

What do you think? What other mindsets or practices have been helpful for you?

May what and how you create enable unimaginable flourishing for you and all life

Do You Have News & Updates?

If you are a team member of a Holochain Project and would like to share your latest news and updates, please…

CONTACT US

If you have a question, request, or simply a desire to connect more about hAppenings.community please click the link above!

The Problem with the “Sharing Economy”

Remember when the sharing economy was supposed to change everything? The pitch was compelling: people would share their spare rooms, their cars, their tools, and communities would become more connected and resourceful in the process.

What actually happened was something quite different.

These platforms captured the language of sharing while building some of the most centralized, extractive business models of the digital era. They sit between peers, control the rules of engagement, own all the data, set the prices, and take a cut of every transaction. The people who provide the actual value (their cars, their homes, their labor) have no say in how the platform operates and no ownership of the relationships they build through it.

This is not a sharing economy. It is a micro-service economy with better marketing.

The problem is not that peer-to-peer coordination doesn’t work. Millions of people successfully coordinate through these platforms every day. The problem is who controls the coordination layer. When a single company owns the rules, the matching algorithm, the reputation system, and the communication channel, “peer-to-peer” becomes a polite fiction. The platform is the real intermediary, and it captures the lion’s share of the value.

So the question becomes: what would a genuine peer-sharing protocol look like? One where the governance of shared resources is not imposed by a platform operator, but embedded in the resources themselves? Where organizations and individuals can coordinate across boundaries without surrendering control to a central authority?

That is the question Nondominium is designed to answer.

What It Actually Looks Like in Practice

Before diving into the history and architecture, consider a concrete scenario.

A makerspace in Montreal has a 3D printer sitting idle three days a week. A cooperative across town needs one for a short production run. Today, they would probably never find each other, or if they did, it would happen through word of mouth with no shared protocol for managing the arrangement.

With Nondominium, that 3D printer exists in a shared network as a resource with a custodian (the makerspace), a governance profile (who can request it, what conditions apply, what maintenance obligations come with use), and a full history of previous custodianship transfers. The cooperative discovers it through the network, initiates a request that triggers the governance rules embedded in the resource specification, and a custodianship transfer is validated and recorded. Both parties receive cryptographically signed participation receipts documenting the interaction. When the production run is done, the machine returns to the makerspace with its maintenance log updated and its reputation trail extended.

No platform takes a cut. No central authority imposed the rules. The governance traveled with the resource itself.

That is the core idea. Now let’s look at where it comes from.

A Decade of Foundations

Nondominium does not emerge from thin air. It is the latest chapter in a story that spans more than a decade of practical experimentation, standard-setting, and software development at the intersection of commons-based peer production and distributed systems.

Sensorica was co-founded in Montreal Quebec, in 2011 by Tiberius Brastaviceanu as an Open Value Network (OVN): one of the first organizations to seriously attempt commons-based peer production outside of software. No employees, no bosses, no fixed hierarchy. Contributors self-organize around projects, pool resources, and share benefits according to transparent contribution accounting.

Running an organization like this surfaces coordination challenges that most economic software simply ignores. How do you track who contributed what when there is no payroll? How do you manage shared equipment when there is no single owner? How do you coordinate resource flows across a network of autonomous contributors without a central authority dictating terms?

These were not theoretical questions. They were daily operational realities.

In 2013, Lynn Foster and Bob Haugen of Mikorzial Software began collaborating closely with Sensorica to build the tools this new form of organization desperately needed. The result was the NRP/CAS (Network Resource Planning and Contribution Accounting System): the first serious attempt at software designed specifically for Open Value Network coordination. It was functional, and it was groundbreaking. But it was also a centralized web application backed by a single database, which created a structural contradiction for an organization built on decentralized principles. Whoever hosted the server ultimately held power over the network’s data and operations.

Despite this limitation, years of running a real peer production network on real software produced hard-won insights into what works, what breaks, and what patterns keep recurring regardless of the specific organization.

Lynn and Bob then recognized that these patterns were universal. Drawing on William E. McCarthy’s REA accounting model (a 1982 framework that describes economics in terms of Resources, Events, and Agents rather than abstract ledger categories), they co-developed the Valueflows specification: a shared vocabulary for describing economic activity in distributed networks, designed from the ground up to coordinate across organizational boundaries.

Valueflows is not software. It is a common language that different applications can speak, allowing them to interoperate without sharing a database or agreeing on implementation details.

The next step was hREA: a Holochain implementation of Valueflows, providing a reference that any Holochain application can build on.

Nondominium picks up this thread and pushes it into specific, uncharted territory: applying Valueflows to the challenge of resource mutualization across organizational boundaries.

How Nondominium Works

Custodianship, Not Ownership

The most fundamental shift Nondominium makes is conceptual.

In conventional systems, resources are owned. Ownership means exclusive control: I own this, you do not, and if you want to use it, we need a transaction that transfers ownership or grants a temporary license.

Nondominium replaces ownership with custodianship. A resource exists in a commons. Specific agents are custodians: they are responsible for the resource, accountable for its condition, and authorized to manage its use according to agreed-upon rules. Custodianship can transfer from one agent to another, and each transfer is recorded, validated, and governed by the protocol.

This is not the absence of property relations. It is a different topology of them. Resources still have stewards. There are still clear rights and responsibilities. But the resource is not locked inside one organization’s database, and no single entity extracts rent from the coordination process.

Governance Travels with the Resource

In Uber, governance is imposed by the platform. In a traditional ERP, governance is implicit in the organization’s internal policies, invisible to anyone outside. In Nondominium, governance is embedded directly in the resources themselves through Resource Specifications. Each type of resource carries its own governance profile: who can become a custodian, what conditions must be met for transfer, what maintenance obligations exist, and what happens in case of damage or dispute.

In practice,: an agent issues an intent to interact with a specific resource. The governance layer checks that all conditions are met and paves the way for the agent to initiate the process. As the agent proceeds, the governance layer modifies the resource’s state accordingly (marking it from “available” to “unavailable,” associating a new custodian, updating maintenance obligations). When a resource moves from Organization A to Organization B, the governance follows. Both parties operate under the same rules, transparently, without needing to negotiate from scratch or trust a third party to enforce the agreement.

Private Participation Receipts

Trust in a decentralized system cannot depend on a central reputation authority like Uber’s star ratings or Airbnb’s review system.

Nondominium introduces Private Participation Receipts (PPRs): every interaction between agents generates a pair of cryptographically signed receipts, one for each party. These bi-directional receipts create an immutable audit trail of resource flows, build verifiable reputation over time based on actual interactions, and function as credentials. An agent who has successfully managed dozens of custodianship transfers for heavy equipment has a provable track record, not because a platform says so, but because the cryptographic receipts exist.

Critically, agents control their own receipts. There is no public scoreboard and no centralized database of everyone’s history. Reputation is portable and agent-controlled.

The ERP Bridge

Because Nondominium is protocol-based, it also functions as an interoperability layer. Organizations do not need to abandon their existing tools.

Through a bridge service currently under development, organizations can selectively publish inventory from their existing ERP systems (Dolibarr, ERPLibre, NextERP) to the Nondominium network, discover resources from other organizations, and coordinate sharing workflows while maintaining full sovereignty over their internal data. The ERP continues to manage day-to-day operations; Nondominium provides the shared substrate underneath.

This positions Nondominium as a bridge between business-to-business and community-to-community resource flows, something that neither traditional ERPs nor consumer-facing sharing platforms are designed to handle.

Why Holochain: The Mycelium Beneath the Forest

The choice of Holochain as the infrastructure for Nondominium is not incidental. It reflects a deep alignment between the protocol’s values and the platform’s architecture.

Holochain is agent-centric: each participant maintains their own source chain of data and interactions. You do not need the entire network to agree on the state of the world before you can act. You maintain your own history, sign your own data, and share it with the network on your own terms.

For commons-based resource coordination, this is exactly right. Organizations participating in a mutualization network do not want to expose all their internal operations to a global ledger. They want to share what they choose to share, maintain sovereignty over their own data, and participate in coordination without surrendering autonomy.

Holochain’s Distributed Hash Table functions as a layer of shared awareness. Resources, their specifications, their governance profiles, and their custodianship histories are discoverable across the network without being stored on any central server.

The metaphor that keeps recurring in this work is the mycelium network beneath a forest. The trees (organizations) appear separate above ground, but underground they are connected by a vast network of fungal threads that allow them to share nutrients, send signals, and coordinate responses to changing conditions. Holochain’s DHT is the mycelium for resource mutualization: invisible, decentralized, and essential for the health of the whole ecosystem.

Holochain also does not require global consensus. Traditional blockchains need the entire network to agree before any transaction is valid. Holochain uses local validation: each node validates the data it receives according to shared rules, and the network achieves eventual consistency through gossip. This mirrors how real-world resource sharing actually works between autonomous organizations. When a cooperative lends equipment to a partner, they validate it locally, according to shared rules, and the information propagates through the network as needed.

The Bigger Picture: Building a Commons Economic OS

Nondominium does not exist in isolation. It is one component of a larger emerging stack for commons-based economic infrastructure.

True Commons (currently in development) is the broader platform for commons-based peer production, co-developed by Tibi and Sacha Pignot as the next-generation NRP/CAS built on Holochain and Valueflows. Where Nondominium handles resources (what exists, who stewards it, how it flows), True Commons handles contributions (who did what, how benefits are shared, what to work on next).

The Requests and Offers application (currently in Alpha testing) from the hAppenings Community addresses the discovery and matching layer: how agents signal what they need and what they can provide. Also built on Holochain and Valueflows, it could serve as a natural front door to the resource-sharing workflows that Nondominium manages behind the scenes.

And Unyt Accounting (a Holochain Foundation subsidiary) provides mutual credit currency infrastructure, enabling high-volume microtransactions without consensus bottlenecks. Integrating mutual credit currencies into Nondominium’s resource flows, so custodianship transfers and service contributions carry economic weight natively within the protocol, is a direction actively being explored.

The vision is composable: Nondominium for resource governance, True Commons for contribution accounting, and Unyt-powered currencies for the transactional fabric connecting them. Each piece remains independent, but together they could form a complete economic operating system for the commons.

The sharing economy promised that technology would help people share resources more effectively. That promise was captured by platforms that turned sharing into extraction. Nondominium is part of the effort to fulfill the original promise, one protocol at a time.

Where We Are and How to Get Involved

Nondominium is currently in active development. Tibi leads the domain architecture and OVN design, drawing on over a decade of experience running Sensorica as a commons-based peer production network. Sacha Pignot handles the technical development, building the Holochain zomes and protocol infrastructure. The core architecture is defined and the MVP focuses on three foundational zomes: person management (agent identity and roles), resource tracking (specifications, lifecycle, custodianship), and governance (state transitions, validation rules, access control).

We are building this as open infrastructure for the commons economy, and we need collaborators. The full documentation index is a good starting point for understanding the project in depth.

• If you are a Rust/Holochain developer interested in commons infrastructure, the codebase is open, and we welcome contributions. Building the economic coordination layer for peer-to-peer resource sharing is challenging, rewarding work.

• If you represent an organization that manages shared resources, whether a cooperative, a makerspace, a tool library, or any network where resources flow between autonomous participants, we are looking for pilot partners who want to test real-world mutualization workflows.

• If you are interested in funding commons-based infrastructure, we welcome your support. Nondominium is positioned at a critical intersection: Valueflows compliance ensures interoperability, Holochain provides the distributed substrate, and the protocol addresses a real, unsolved problem in the cooperative and commons economy.

The sharing economy promised that technology would help people share resources more effectively. That promise was captured by platforms that turned sharing into extraction. Nondominium is part of the effort to fulfill the original promise, building the tools for genuine resource mutualization, one protocol at a time.

Want the Full Deep-Dive?

The complete article covers:

• The full intellectual lineage from McCarthy’s 1982 REA model through Valueflows to Nondominium

• The multi-layered validation architecture (semantic, cryptographic, and social)

• How to get involved as a Rust/Holochain developer, pilot partner, or funder

Read the full article on Alternef Digital Garden

What resources sit idle in your network? Have you ever tried to share them across organizational boundaries, and what broke down when you tried? I’d love to hear what coordination challenges you’ve run into.

Connect with the Sensorica Team on the Discord Server!

Useful Links

Nondominium
Sensorica
Sensorica GitHub
hREA
ValueFlows

If this resonated, subscribe to Alternef for more explorations at the intersection of commons economics, distributed systems, and alternative economic infrastructure.

🌟 Latest News from hAppenings.community