One AI gives you an answer. Two AIs fighting give you the truth.
GPT-4o vs Claude vs Gemini vs Llama. 19 providers. 6 fight modes. No mercy.
You wouldn't make a million-dollar decision based on one person's opinion. Why do it with one AI? Make them fight.
19 providers · 6 fight modes · SSE streaming
Not a coding assistant. A programmable AI workforce. You design the pipeline. You pick which AI brain handles each step. Same tool, any role.
Stress-test contracts, analyze threats, find regulatory gaps before they cost real money
War-game pricing, positioning, and go-to-market plans before committing budget
Debate architecture decisions, review code approaches, run adversarial security audits
Bull vs bear on any asset. Stress-test financial models before reality does
Pre-submission peer review, debate prep, hypothesis testing from every angle
A/B test ad copy, debate editorial direction, refine UX writing before launch
Six distinct ways for AI models to clash. Pick a mode, set the rules, and let them fight.
Two models debate head-to-head on a topic
1v1 with a third model scoring and judging each round
Multiple models clash in an open arena
Models build on each other's arguments sequentially
Bracket-style elimination across multiple rounds
A panel of models discusses and deliberates together
Add API keys for any of the 19 supported providers. Keys are encrypted in your personal vault with AES-256-GCM.
Set a topic, upload reference documents, pick a mode, and watch models battle in real-time via SSE streaming.
Generate AI summaries and detailed PDF reports of completed fights. Share transcripts on the public feed.
API keys encrypted per-user with scrypt-derived keys. Password never stored on server.
Attach reference documents to fights. Models debate using your source material.
Generate AI-powered fight summaries and detailed PDF reports with key arguments.
Connect external knowledge bases so models argue with real data, not just training data.
Free to use. Bring your own API keys. Watch the models destroy each other's arguments in real-time.