One place for chorus coding: several sessions of Claude Code, Codex, Gemini CLI, Kimi and Antigravity working together, with you in charge. Real terminals, one Inbox for everything they ask you, and a morning page that says what happened overnight. Free and open source.
I usually have five to ten coding agents going at once: Claude Code on the careful work, Codex reviewing it, Antigravity on routine changes, and loops that re-run tests overnight. I call this chorus coding: several agents on the same work, each doing the part it does best, one checking another.
The agents stopped being the bottleneck. I did. Questions sat in tabs I wasn't looking at, nothing told me what failed at 3 AM, and every terminal looked the same.
One Roof is what I built to keep up:
- Inbox: every question from every session in one place, with the agent's own options as buttons. Nothing is approved for you.
- Home: what ran overnight, what broke, and what to do next.
- Projects: plan in Linear, run the work here with many agents, results written back.
- Resources: what each session and app is using, with buttons to act on it.
- Loops: about 25 ready-made loops that watch, check and report, with limits and gates.
Each session stays the real agent in a real terminal, so you can always type straight into it.
It's free and open source (MIT). For now it runs on macOS 14+ on Apple silicon. Evals is still a draft, and I'm the only one who has used it day to day, so I'd love to hear what breaks or reads badly.
the linear part caught my eye. when two agents end up reading the same ticket differently, does that come back to the inbox as a question, or do you only find out at review?
Our worst night here was an evening routine that kept reading feeds until the morning and never wrote a thing, and nothing anywhere showed an error. Would the morning page flag a run like that, or does a run that never finishes just look quiet?
An agent waiting for an answer and a failed tool call need different attention. Keeping those separate in the morning view would make it easier to decide where to start.
About One Roof on Product Hunt
“A new way to build with agents. You stay in charge.”
One Roof launched on Product Hunt on November 3rd, 2026 and earned 79 upvotes and 8 comments, placing #20 on the daily leaderboard. One place for chorus coding: several sessions of Claude Code, Codex, Gemini CLI, Kimi and Antigravity working together, with you in charge. Real terminals, one Inbox for everything they ask you, and a morning page that says what happened overnight. Free and open source.
One Roof was featured in Open Source (69k followers), Developer Tools (522.4k followers), Artificial Intelligence (482.8k followers) and GitHub (41.5k followers) on Product Hunt. Together, these topics include over 278.1k products, making this a competitive space to launch in.
Who hunted One Roof?
One Roof was hunted by Gaurav Narasimhan. A “hunter” on Product Hunt is the community member who submits a product to the platform — uploading the images, the link, and tagging the makers behind it. Hunters typically write the first comment explaining why a product is worth attention, and their followers are notified the moment they post. Around 79% of featured launches on Product Hunt are self-hunted by their makers, but a well-known hunter still acts as a signal of quality to the rest of the community. See the full all-time top hunters leaderboard to discover who is shaping the Product Hunt ecosystem.
Want to see how One Roof stacked up against nearby launches in real time? Check out the live launch dashboard for upvote speed charts, proximity comparisons, and more analytics.
Gaurav Narasimhan
Hi Product Hunt! I'm Gaurav.
I usually have five to ten coding agents going at once: Claude Code on the careful work, Codex reviewing it, Antigravity on routine changes, and loops that re-run tests overnight. I call this chorus coding: several agents on the same work, each doing the part it does best, one checking another.
The agents stopped being the bottleneck. I did. Questions sat in tabs I wasn't looking at, nothing told me what failed at 3 AM, and every terminal looked the same.
One Roof is what I built to keep up:
- Inbox: every question from every session in one place, with the agent's own options as buttons. Nothing is approved for you.
- Home: what ran overnight, what broke, and what to do next.
- Projects: plan in Linear, run the work here with many agents, results written back.
- Resources: what each session and app is using, with buttons to act on it.
- Loops: about 25 ready-made loops that watch, check and report, with limits and gates.
Each session stays the real agent in a real terminal, so you can always type straight into it.
It's free and open source (MIT). For now it runs on macOS 14+ on Apple silicon. Evals is still a draft, and I'm the only one who has used it day to day, so I'd love to hear what breaks or reads badly.
The longer why: https://console.zibajoon.com/blog/chorus-coding