AskVerdict
AI agents debate your decisions so you don't decide alone.
I kept running into the same problem. Ask ChatGPT a hard question, get one confident answer. Ask it again differently, get a slightly different confident answer. Neither tells you what it's leaving out.
So I built a system where AI agents take opposing sides on your question. They research, argue, cross-examine each other, and produce a structured verdict with confidence scores and dissenting views.
Think of it like a boardroom of advisors that actually disagree with each other, for under a dollar.
The stack: TypeScript monorepo, Hono + Bun backend, Next.js frontend, Postgres, Stripe. BYOK model - users bring their own API keys for Claude, GPT, DeepSeek, or Groq. We don't store keys, don't pay for compute. Users pay a subscription for the platform.
Unit economics: zero marginal cost per debate since users supply the API keys.
Revenue comes from subscriptions starting at $3.99/mo.
Site is live at askverdict.ai. Opening to the first 100 users soon.
Curious what this community thinks — would you use something like this? What decisions would you test it on?
Comment