I've also updated the Proof of Concept on the site here, and fixed the wasm-opt issue that made the binary there roughly 1/3 larger than it needed to be. Yay!

Since this seems to be in a somewhat ongoing support mode, at least while it's Summer Camp season and people are actually working on things upstream, I'll do this periodically to keep up compatibility. The goal there is still an eventual green field rewrite on top of the Maypole primitives and protocols, but that's still a ways out.

Since it's in semi-official support, I'm rolling builds for Desktop (Linux/Mac) and 64-bit Android. These should be available on Codeberg within a few days (need to get home and do the Mac build).

We've also got a new Rosie icon and header by Andy Carolan and I'm super excited to start rebranding around them!

I have 4.5 days work on this, give or take day job and farm responsibilities which are fairly light at this moment. This is the end of day 3, and there has been a huge amount of progress!

Astod is nearly feature complete (if ugly). Identities can be created and secured, Personas can be created and destroyed. Apps can request permission to use Personas, and can have their Persona list updated or all permissions revoked. This works on Linux, Mac OS, and... within a few minutes... Android.

As a test app I'm building out a simple poll / survey app that allows people to have simple interactions at the Persona level. Astod is written in Rust using Tauri, which has an excellent plugin for running cross-plarform services. Maypolls is written in Dioxus, to make sure Tauri doesn't leak into the crates that other apps use.

As of today Maypolls can complete the whole Persona-App lifecycle on Linux and Mac OS, but can't actually create or work with polls! The next 1.5 days will be trying to get basic polls working, but then we need friends to share them with and invitations to grant them permission to view them. The 'Contacts and Trust' and 'Invitations' protocols aren't even written up yet, but I may just write up quick versions to prove out the concepts while there is time.

Check out the code and the protocol documents at the Codeberg org, feel free to offer feedback or join the team to help try to get Veilid into more hands.

So the first thing I wanted to work on when writing protocols was finding a set of functionality that people expect and mapping those concepts to the way things work in Veilid. It was, honestly, a pain in the ass. If work together to get it right, though, fewer people will have to do it in the future.

The protocol attempts to cover the lifecycle that users expect, offer the best security and privacy guarantees that I can, and make multi-account, multi-device accounts seamless.

So take a look and create some issues to help us build the blocks that can help build a better network!

He has always been motivated to work on communication tools, hacking on IRC bots in the 90s and developing chat, voice chat, message boards, leaderboards, and many other social tools for console and PC games. When learning a new language or technology, he writes chat. So far he has implemented some form of chat app in C, Perl, SHTML, PHP, Java, C#, Cold Fusion, C++, Python, Go, Objective C, Dart, and Rust.

And so it came to pass that Veilid was announced at Def Con in 2023, and Dumont played around with it and the associated Veilid Chat PoC that was announced at the time. There were still some rough edges, but there was a real sense that this would be The One... a technology that unlocked 30 years of ambition for what chat could be.

The events of 2025 and growing surveillance in both business and public life rekindled the spark of wanting to build a peer-to-peer, private chat app that anyone could put power back into the hands of the user. There didn't seem to be much progress on Veilid or Veilid Chat, and so he started working on a protocol breaking XMPP client that would use newer encryption and MLS along with the Spaces features being developed in Movim.

Right around the time Dumont got a client running he learned of Scrapwire on Mastodon, which provided the building blocks for just what he wanted to build, and a whole lot more. It appeared to be significantly AI written, which coincided with a mandate from his day job to learn all about LLMs first hand and 'form opinions'. So Maypole was born of the sudden ability to create a full-featured chat app on Veilid in short order, iterate quickly, and evaluate the feasibility of building such a tool for real.

Just in time for May Day 2026 a usable proof-of-concept was uploaded here. Given the quality of the backing software Dumont didn't want to release all of the builds, but a WASM version built with deep hacks into Scrapwire was made available on this site. There are also more fully functional builds working for Android, iOS, Linux desktop, and Mac OS.

We are now kicking off a new phase: to write a new protocol that has somewhat different priorities from Scrapwire and rounds off some of the rough edges, while building on some of the foundations such as MLS. We are also working on related protocol for Veilid-based identity and reputation systems. Then will come Rust crates implementing the protocols and abstracting the Veilid layer. Finally Maypole itself will be built from the ground up with whatver the best technology is at the time.