screenpipe

Tech stack & purpose

Screenpipe is a local-first screen and audio recording application that captures accessibility trees and conversations for AI agent consumption. Built with Tauri, Rust, TypeScript, and Bun, it serves as a context layer for agents by indexing captured data locally for private, on-device processing. The project is backed by Y Combinator (S26) and includes extensive end-to-end testing infrastructure across Windows, macOS, and Linux, along with specialized testing for features like HD recording pipelines, audio fallback scenarios, and agent integrations with systems like OpenClaw and Hermes.

Community & reference links

• discord.gg/screenpipe
• docs.screenpi.pe
• docs.screenpi.pe
• OpenClaw integration guide
• github.com/screenpipe/screenpipe/issues
• #3466
• #3107
• Streamable HTTP transport
• alpha version of our AI PII model
• tart
• www.pulsemcp.com/servers/screenpipe-screenpipe
• www.reddit.com/r/screen_pipe
• www.youtube.com/@screen_pipe

Languages

Contributors (top 30 of 176)

• Louis Beaumont9,803 commits
• matthew-heartful331 commits
• Ezra Ellette320 commits
• GitHub Actions Bot218 commits
• Ansh Grover195 commits
• Divanshu148 commits
• louis030195122 commits
• Harsh Kumar119 commits
• Matthew Diakonov88 commits
• Neptune70 commits
• mintlify[bot]66 commits
• Anshgrover2354 commits
• Divanshu Grover45 commits
• shaik-zeeshan45 commits
• tribhuwan41 commits
• neo77340 commits
• screenpipe-agent[mac-mini]33 commits
• Joe Goldin27 commits
• Vitalik24 commits
• Chirag Singhal21 commits
• AI Agent20 commits
• cparish31218 commits
• ologbonowiwi17 commits
• tribhuwan-kumar14 commits
• Wesley Matos13 commits
• divanshu-go13 commits
• thisisharsh712 commits
• oliver10 commits
• kerosina9 commits
• Isaac Sanchez-Hawkins8 commits