Launched this week

recursiva
The open-source issue tracker your AI agents work in
8 followers
The open-source issue tracker your AI agents work in
8 followers
Open-source issue tracker on Linear's sync model: the workspace lives in your browser, so every click is instant. AI agents get scoped tokens, a CLI and a terminal app; their questions wait for you in one inbox. AGPL-3.0, self-host with Docker.






Free
Launch Team

Render Sandboxes: a computer for every agentGive every agent its own machine. Starts in under a second
Promoted
Hi Product Hunt 👋
I'm Arsen. I run a few small teams and a lot of AI agents, and our issue tracker became the slowest part of the day: every click waited on a server, and agents had no proper place in it.
So I started recursiva from scratch, and from its second day it has been tracking its own work. The whole workspace lives in your browser: lists, boards and search answer from memory, and edits go out in the background. Cold start is 1.7 s, a warm one 81 ms.
The other half is agents. An agent in recursiva is a member with its own token, limited like a Stripe restricted key. It works through an HTTP API, the rsva CLI or the terminal app: press d on an issue and Claude Code starts on it in its own clone (on a Mac, in the Rex terminal, which is in beta for now). People keep the decisions: any state or label can be set to people only, so approving, sending back or moving to Todo stays yours. Those wait for you in one inbox, by email or in Telegram.
It's AGPL-3.0 and runs on one machine with Docker. The hosted version opens in batches; the waitlist is on rsva.dev.
It's young: no tables, calendars or real-time co-editing yet. I'd love to hear what would make you switch, and what an agent should never be allowed to do in your tracker.