The Clubhouse / Twitter Spaces craze may have come and gone, but that doesn’t mean the ‘verse can’t have their own version of it. Audon is a simple multi-user voice chat app that’s fun, easy to use, and reliable. Best of all, the app relies on some neat tricks to make the experience seamless.

A Different Approach to Apps

Audon differs from other Mastodon apps in that it’s not much of a client at all. Users sign in with a Fediverse account, a session grabs their details, and a temporary link to a shared space is created for the two to chat. As long as your platform of choice supports the Mastodon API, it’s possible to use a space.

The architecture choices in the app seem like a good fit, too. Audio is streamed over a WebRTC connection and a STUN server, with Redis doing the heavy lifting for peers. The main Audio component appears to be powered by LiveKit, which aims to keep everything as smooth as possible for developers.

Damon and I talked for upwards of an hour, with no noticeable drop in call quality within the space. Although the UI is a bit simple, it’s incredibly easy to use. During our time, we noticed no call drops, quality degradation, or latency issues. Granted, that situation might be different with a few dozen people in the space, but our initial test was promising.

Slightly Disappointed

One surprise came from realizing that Audon is actually no longer in development. The source code is publicly available under the GNU AGPLv3. If anybody wants to fork it, modify it, or otherwise study it, Audon could provide some great lessons to build on.

The Good:

• The interface and onboarding process are super, super easy to work with.
• The hosted instance is simple and reliable.
• Supports Mastodon, Pleroma, Firefish, and anything else with a good Mastodon API implementation?
• Treating this as a Mastodon client app, and not a full-fledged Fediverse platform kept the entire process simple.
• Call quality for two people was fantastic the entire hour and a half that we chatted.
• Emoji Support!

The Bad:

• I clicked on some emoji and couldn’t really see anything visually happen.
• Lack of an input/output switcher feels like an oversight.
• The app is a little bare-bones, and feels a bit more like a demo?

The Ugly:

• While source code is available under a FOSS license, development appears to have stalled. The main repo is archived, with almost no development over the past four months.
• You have to user Docker to self-host it. Not the worst thing in the world, but some people *hate* Docker.

 Share