Start with the track, not the logo
A famous host helps, but the track decides what you can actually build. Look for prompts that match your data access, teammate skills and ability to demo a real workflow.
AI hackathons with agents, evals, retrieval, robotics, biotech or video tracks all reward different teams. Pick the track where your unfair advantage is visible.
Check whether the deadline fits your life
A seven-day online sprint and a forty-eight-hour onsite event are different games. If your team is remote, avoid formats that require constant synchronous work unless everyone can commit.
The best deadline is close enough to create urgency and far enough away to test the demo with one real user before submission.
Compare prize money against friction
A $100,000 prize pool can be worth it, but only if the rules are clear and the track is not overcrowded. Smaller events often give better odds and more useful sponsor feedback.
Read eligibility, payout timing, IP terms and required APIs before joining. Bad terms can make a large prize less valuable than a clean smaller event.
Use the index as a shortlist
Use Tensor Hack to find live AI hackathons, then open the source page for the two or three that match your team. Do not enter everything.
A focused shortlist gives you time to form a team, validate the prompt and build a demo that is judged on substance rather than luck.
< read by a human · updated as things change >
browse hackathons