I built an AI that doesn't agree with you — it convenes a council

You know the feeling. You ask an AI a hard question, it gives you a confident answer, you act on it, and later you realize it was wrong — but there was nobody in the conversation to push back.

That's the problem I built Deliberon to solve.

It's not a chatbot. It's a council chamber. You bring a question to a Chairman who assembles a team of specialist agents — Advocate, Adversary, Skeptic, Strategist, Risk Analyst — each with a distinct role and perspective. They deliberate in structured rounds. Anti-sycophancy detection flags any agent that shifts position without citing new evidence. Minority positions are preserved in the final Decision Record. Every session is cryptographically signed.

The result isn't a chat reply. It's pressure-tested reasoning you can actually use — from a team you convene but don't have to hire.

Right now it's a free-to-start Windows desktop app. Bring your own models — local, cloud, or terminal agents. Paid tiers are still being finalized, so nothing on that front to share yet.

I've been building this solo for a while and the product page just went live. What I'd love to hear from this community: what decisions do you wish you'd stress-tested before you shipped them? The kind where having a room full of specialists to argue it out would have saved you.