roundtable.audio (formerly discourse.fm) is a web app for hosting live, moderator-decentralized audio chat rooms (or "roundtables"). These can be public or private to join or listen to, and host up to 100 participants.
While it's possible to start ad-hoc roundtables about anything, part of what differentiates roundtable.audio is integrating external sources to populate discussion topics. Currently, the only sources are:
- Twitter: you can start a roundtable by tweeting at the roundtable Twitter bot from any tweet thread and including the words "start a roundtable." Give it a try and see what happens!
- Y Combinator's Hacker News: For the Hacker News integration, you can visit hackernews.roundtable.audio to see a page which intends to emulate the Hacker News homepage, but rather than having a comment section for each story, you can pop into a roundtable. I'd like to eventually have sources for subreddits, popular news sites, other web-based communities, and more. Wherever people are synchronously talking via text on the web, it can probably benefit from the addition of roundtables/live audio.
This codebase is far from a polished product, and it would be great to have others get involved - especially on the addition of more sources and the cultivation of their adoption. If you'd like to help/hack on any aspects of this project, send an email to [email protected] or open issues/pull requests in this repo. It would be greatly appreciated!
This repo is organized into a number of microservices running on Docker. For ease of understanding, it's simplest to just explain what the notable ones do.
Handles the serving of client code from an nginx proxied server. All of the frontend code can be found in the client-code directory. Take a look at some of the neat visualization modules in found in the js subdirectory!
Controls the core management of roundtables, including authenticating user credentials (for private roundtables), passing of WebRTC credentials, chat control, and more. Essentially, it's the backend that controls everything except for the relaying of audio data.
This is the core audio relaying service, which is an untracked clone of a feature I built on top of the Kraken SFU. The feature branch can be found here. This SFU uses a Golang implementation of WebRTC built by the Pion community. Many thanks to these open-source projects for the extraordinary work they've done in enabling new possibilities with WebRTC.
This is a simple implementation of the open-source coturn server. A TURN server is required for WebRTC signaling. While it's not strictly necessary to self-host one, it's probably the best and most reliable way to go.
Controls the Twitter bot which continuously listens for tweets sent to it containing the right keywords, creates new roundtables, and returns the links to the user.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent