Yoetz - Check to see if AI really completed a task or if it had gaps

Hey Product Hunt! 👋 I’m Shay, building Yoetz with my cofounder Daniel. I was using coding agents to compile a large Nasdaq-related database for my university research. They could do a lot of the work, but I kept finding small mistakes and instructions that hadn’t been followed. I wanted to focus on the research, not keep checking whether the agent had done everything I asked. Daniel was running into similar problems with his studies. We both saw how powerful these tools were. We wanted to help them follow through, without having to keep nudging them ourselves. That’s why we built Yoetz. Yoetz isn’t another coding agent. It works alongside yours, using tools and supported hooks to record work evidence and check what supports the agent’s claims. It combines deterministic checks with optional model-powered review, then returns findings the agent can address. For example: the agent says “tests pass,” but the recorded test run happened before its latest code change. Yoetz can flag that stale evidence so the agent can address it. It doesn’t guarantee correct code. It shows what was checked, what wasn’t, and what still needs attention. Yoetz is open source and local-first, with integrations for Codex, Claude Code, and Cursor. We’d love you to try it on a real task and tell us what it catches, what it misses, or where it gets in your way. What do you still find yourself checking after your coding agent says it’s done?