Screenpipe YC S26: Record Workflows to Power AI Agents

Screenpipe, a Y Combinator S26 startup, has launched a tool that records on-screen activity and converts it into AI agents, exposed via an API. The client captures mouse movements, clicks, keyboard input, and UI context from any application, aiming to lower the barrier for creating agents that interact with existing software. The launch has sparked discussion on Hacker News.

Screenpipe, a startup that closed a Series A round as part of Y Combinator’s S26 cohort, unveiled a tool that records on‑screen activity and turns those recordings into AI agents. The client captures mouse movements, clicks, keyboard input, and UI context from any application, stitches the data into a reusable agent, and exposes it via an API. The launch sparked discussion on Hacker News, where users noted the potential to lower the barrier for creating agents that interact with existing software. Q: Can Screenpipe work with web applications that require authentication? A: Yes. The recorder captures login flows, and the generated agent can replay those steps, handling cookies and tokens. Q: How does the agent handle dynamic content or pop‑ups? A: The recorder logs timing and context of dynamic elements. The agent includes logic to wait for or dismiss pop‑ups based on the recorded behavior. Q: Is the tool suitable for production use, or is it still experimental? A: Screenpipe is already used in pilot projects by several teams. While it’s still maturing, the core functionality—record‑to‑agent conversion—is stable enough for production workloads that don’t require strict real‑time guarantees. Originally published on Automations Cookbook.