Ask a hard question. AskVerdict spawns AI agents with opposing stances that research, argue, cross-examine each other, and deliver a structured verdict with confidence scores and dissenting views. Not a chatbot giving you one safe answer. A debate engine that pressure-tests your decisions. Multiple AI providers. Bring your own API keys. Three modes: fast, balanced, thorough. Built at Anthropic's Claude Code Hackathon, now a full product. First 100 beta users get early access

Hey Product Hunt! I'm Gagan Deep, founder of AskVerdict. This started at Anthropic's Claude Code Hackathon. 14,000 applied, 500 got in, I had one week to build something. The problem was simple: when you ask an LLM a hard question, you get one answer, one angle, zero pushback. So I built a system where AI agents with distinct perspectives actually debate each other. Forced cross-examination. Structured argument tracking. A verdict with confidence scores and dissenting views. A doctor tested it during the hackathon on a medical question physicians genuinely disagree on. ChatGPT gave her the same answer fifteen times. AskVerdict gave her honest tradeoffs in five minutes for under fifty cents. We didn't win the hackathon. But the idea proved itself. The prototype became a full product. We're opening up to the first 100 beta users. Would love your feedback. What decisions would you debate first?

Quick update since launch: we've shipped table rendering in debate arguments, a stage progress bar so you can track where the debate is, and smart auto-scroll that follows the verdict. Building in public, shipping fast.

https://askverdict.ai/

AskVerdict

AI agents debate your decisions so you don't decide alone.

AI agents debate your decisions so you don't decide alone.