Rerun
The easiest way to build AI agents for all your tasks
748 followers
The easiest way to build AI agents for all your tasks
748 followers
Most AI agents are black boxes. Rerun isn't. Build no-code agents that run 24/7, chasing invoices, qualifying leads, clearing your inbox, and watch every step in real time. They pause for approval before anything sensitive. Each workspace gets its own private server. Start with the included model, or connect your own API key or subscription
This is the 2nd launch from Rerun. View more
Rerun
Launching today
Rerun turns your ChatGPT or Claude subscription into a team of AI coworkers. Each one has tools, a computer and a shared workspace, runs on a schedule or a trigger, and asks before risky actions. Now open-source (AGPL v3) and self-hostable.
Free Options
Launch Team / Built With
AWS Activate CreditsCredits that fuel your startup’s next milestone
Promoted
Hey Product Hunt 👋
I'm Clément, and with Théo we build Rerun: AI coworkers that actually do the work. Each one gets its own tools, its own computer (a real browser you can watch and take over), a memory and a schedule. They chase invoices, qualify leads and sort your inbox, and they stop and ask before they send, pay or delete anything.
Today Rerun goes open source under AGPL-3.0. I've been a developer for 15 years, and almost everything I've built stands on open source. Giving back felt obvious, and Rerun is the most complete product we've ever built, so it's the right one to share.
Clone the repo, run docker compose up, and your coworkers run on your own server, on the Claude or ChatGPT plan you already pay for, or any model through OpenRouter. If you'd rather not run anything, Rerun Cloud is the same code, ready in 30 seconds.
docker compose and the subscription u already have makes the self hosted option easier to consider
i would want the setup guide to make it clear which connections stay local and which still depend on rerun cloud
@saad_el_gueddari Thanks for the feedback, we're working on it
We want to open up even more features that are, for now, still Cloud-only
We needed our partners' approval before we could release publicly, but that's coming in the next few hours
For invoice chasing, can it check whether a payment arrived before sending the next reminder? That would be my first workflow to try.
@vadym_mosesov1 Yes!
@vadym_mosesov1 Yeah of course, just ask a coworker, they'll set everything up for you
congrats on the launch clement. asking before a risky step is the part that makes me trust a coworker agent. soft wording on the prompt is what burned me before, so i hope it shows the exact action not only a nice summary
Open sourcing under AGPL is a bold call for something that runs on people's own ChatGPT/Claude subscriptions. When an agent asks for approval before a risky action, does it show the exact payload it is about to send (the email body, the amount), or just a summary of the intent?
@galdayan Thanks for your feedback
It tries to be clear and concise by default, but you can ask it to be more or less detailed depending on what you need!
I’d start with sorting emails and drafting replies. The pause before sending would make that a much easier first step.
Congrats on the launch! If I spot a mistake while watching an agent run, can I pause it, adjust a step, and resume from there? That would be useful for debugging longer workflows.
@stan_xu1 Exactly
You can tell your agent how to improve, and it'll update its processes automatically
But it'll never do it without your approval, so it doesn't go off the rails