Launching today
Ombuto Code
From requirements to running code. Autonomously.
1 follower
From requirements to running code. Autonomously.
1 follower
Coding agents write code fast, but someone still has to prompt, wait and check. Ombuto Code takes you out of that loop. It is an open-source desktop workbench that lives in your repo. Write a PRD, architecture and tickets with an agent, each ticket with acceptance criteria. Switch on the scheduler and agents build, test and fix in a loop while an evaluation agent checks the result against the criteria. Review the report, run summary and diff. Works with Claude Code, Codex and Kimi. Apache 2.0.
Hi everyone, I'm Francois, maker of Ombuto Code.
I love creating software with AI coding tools, and I wanted to free up as much of my time during the build as possible (to build more, of course). I realised I had two options: spend my time interactively steering the agent, or frontload the build with the right context and get out of the way.
Ombuto Code is built around the second option. Agents help you define requirements, architecture and features, then break them down into tickets with acceptance criteria. Tickets then run completely autonomously, honouring their dependencies. After each ticket the agent runs the automated tests; any failing test sends it back to the build phase. Once the build passes, a separate evaluation agent checks the outcome against the acceptance criteria. Only then does a human get the final say: accept or reject, with the evaluation report, run summary and full diff in front of you.
The result is that I get far more done in a day than I did interactively prompting, and watching the board move on its own is genuinely fun.
It's free and open source (Apache 2.0) and works with Claude Code, Codex and Kimi. Try it with:
npx create-ombutocode my-project
I'd love to hear if a specs-first approach works for you, and what you'd want the review step to show. Also, any feedback will be appreciated. I'll be here all day.