Cofounder

Most 'AI, rate my idea' tools give the same beige feedback no matter what you paste. Cofounder doesn't copy how these founders talk — it encodes how they decide: the questions they ask first, what they obsess over, what they dismiss, and how heavily they weight each thing. So the same idea genuinely diverges — Brian Chesky scored a dinner-party startup a 67 while Naval gave the exact same pitch a 34.

Every verdict is a SHIP / SKIP / PIVOT with a transparent five-part rubric — wedge, moat, timing, market, founder-fit — where each founder weights the dimensions differently. The killer question, the one reason it dies, and a concrete pivot that saves it all come in their voice. The score is the argument, not a vibe.

The part that makes it more than a critic: you can argue back. The founder defends the verdict and the score moves — but only for real traction, a sharper wedge, or a genuine moat, never flattery. I argued a gym-bot idea from a PIVOT (59) up to a SHIP (74) by showing it was already running a real gym. Every idea posted lands on a public leaderboard, and it tells you if someone already pitched something close. The brainsets channel each founder's public playbook — not affiliated, no fabricated quotes.