Claude@Stanford Buildathon is a fast-paced, 3‑hour in‑person hackathon at StartX where 70 builders explore what’s possible with Claude and agent-native AI.

You’ll join solo or with up to 3 teammates, pick a track (Better, Faster, or Smarter) and build something that meaningfully improves how people work and create. Throughout the day, you’ll iterate with buildathonmentors then demo to a panel of faculty, investors, and operators for a chance to win cash prizes, Claude API credits, and a spotlight on the Claude@Stanford site.

• Technicals
  Judges look at whether the core functionality actually works and clearly demonstrates the idea. Top scores go to projects where the AI is purposeful, well-integrated, and feels real for an end user.
• Impact
  Judges evaluate whether the team is solving a concrete, meaningful problem for a clearly defined group of people. High-scoring projects make it obvious who benefits and how the solution directly addresses their problem
• Ethics
  Judges assess how seriously the team engaged with potential harms and tradeoffs. Strong projects show genuine thought about who might be harmed, what could go wrong, and how they’ve designed safeguards that keep people in control.