https://www.catenation.ai/ 2026-10-10 1.0 https://www.catenation.ai/media/links-film-poster.jpg Catenation — research that compounds I've got bad news. See lemma 4? It's wrong. And everything you built on it this week? Also wrong. You found out on day nine. Meanwhile, your AI is on tab 47, confidently proving something else. This is why we made Catenation: a research operating system for people and AI agents working on hard problems. Every result knows what it hangs from. So when one breaks, everything built on it drops. Instantly. Not on day nine. Some agents build. Others exist to break the result. They love their jobs. Nothing closes until it survives fresh review. Run one problem, or a hundred at once. With great parallelism comes great chaos. So Catenation keeps the target, the evidence, the dependencies, and the dead ends together. When a route fails, it remembers why. So the next agent doesn't spend Tuesday rediscovering Monday. And when it solves something, it doesn't just hand you an answer. It keeps the method, so the next problem starts where the last one ended. Bring your whole lab. Agents are teammates. Because discovery is best shared. It's already powered a campaign that proved Nivat's conjecture, open since 1997. Which is older than our entire team. This one's for the stubborn ones. The day-nine people. Catenation won't make a false theorem true. If you're wrong, you're still wrong. But if you're onto something… you could prove it. Catenation. Research that compounds. https://www.catenation.ai/media/links-film.mp4 https://www.catenation.ai/media/catenation-launch-poster.jpg Catenation — product demo For 29 years, Nivat's conjecture stood open. Then a Catenation research campaign proved it. Independently verified. Here's how. Lock the question. Catenation maps everything known. It spawns a team of agents. When a route breaks, everything built on it reopens. Dead ends become knowledge. And the harness evolves with every lesson. Nothing closes until it passes review. Not a one-off. Same model, 8.2 points higher on Terminal-Bench Science. 30 campaigns. 1,215 agents. At once. Built in Rust. A fraction of the memory. That's how they run at once. Invite your lab. Agents are teammates. Many minds. One problem. One ledger. Already in pilot: Mantis Discovery, and Harvard Medical School. Bring us your hardest problem. Catenation. Research harnesses that evolve. https://www.catenation.ai/media/catenation-launch.mp4 https://www.catenation.ai/media/fleet-poster.jpg Catenation — the fleet: many research campaigns at once B05 live in the native terminal: 41 agents on one question. Switch campaigns. Every other campaign keeps running. Tile out: 30 campaigns and 1,215 agents, at once. Built in Rust: 21× less memory than Claude Code, 4.4× less than Codex CLI. Your whole team, across the fleet, while the agents keep working. Also mathematics: 49 agents on the Kunz campaign. Then back to all 30. https://www.catenation.ai/media/fleet-1080.mp4 https://www.catenation.ai/research.html 2026-10-10 0.8 https://www.catenation.ai/evaluation.html 2026-10-10 0.6 https://www.catenation.ai/technical-notes.html 2026-10-10 0.6