Spacesheep: Documents of the future, made by AI, read by your team

Teams can comment on a specific line, and mentioning the built-in agent in a comment makes it revise the space while everyone watches. Comment on the exact line. Write @spacesheep and the AI fixes the space while everyone watches. Version history keeps every edit. Spaces can also show live data streamed from servers and sensors, and connect to the tools your team already uses. The company's own demonstration pointed an agent at YC's public directory of 6,305 startups across 21 years and returned a narrative split into six eras, from consumer web to agents and robots. Pointing an agent at the directory produced one story in six eras. The CEO previously worked at Anthropic, where he built the internal app deployment framework. The founder has also said that his agents sent him 355 hours of reading this week. The sources I reviewed did not name direct competitors. For founders building agent products, the bottleneck is increasingly human review of agent output rather than generation. A shared surface with comments, version history and search gives a team a place to check and steer agent work, and it can double as a artifact for customers. The product is early, with two employees and a Fall 2026 YC batch listing, so adoption claims should wait for usage data.