Petals: One Brain Infinite Outcomes

Hey Product Hunt! 👋

We built Petals because we were tired of running five AI coding CLIs in five terminal tabs that had no idea the others existed.

Petals runs Claude Code, Codex, Grok, Kimi and Kiro side by side on the same project. The trick is one shared file, PETALS.md, that lives inside your repo. The plan lives there; every agent claims its own slice of the work, they leave each other one-line notes, and they declare DONE there. One prompt becomes a coordinated work distribution instead of five agents stepping on each other's edits.

A few things we care deeply about:

🧠 One shared brain. PETALS.md is plain Markdown in your repo. No cloud sync, no proprietary format, readable by humans in code review.

🔒 Runs entirely on your Mac. Your logins, your local models. Petals never proxies your code or your tokens, and never marks them up.

🌸 Petals Build, the built-in house agent: local models, zero API keys, 30+ slash commands, every prompt checkpointed.

🍎 Native macOS craft: signed and notarised, universal binary, under 5 MB.

Pricing is simple on purpose: every feature in every plan. $45.99 a year, or $99.99 once for a lifetime license.

We would love your feedback, especially on which agent you want us to add next (Gemini CLI is currently winning that argument internally). I will be here all day answering everything.

🌸 https://petals.f-r.co