A curated list of resources, integrations, pipes, and community projects for Screenpipe — AI that knows everything you've seen, said, or heard.
Screenpipe captures screen text and audio locally while recording is enabled, making computer history searchable for people and AI assistants. Its source is available under the Screenpipe Commercial License, not an open-source license. Optional cloud AI, transcription, sync, and connected services can transmit context off-device; see the privacy data flow.
- Official Resources
- Official Libraries
- Built-in Pipes
- MCP Server
- AI Coding Tool Integrations
- Community Projects
- Service Integrations
- Tutorials & Articles
- Screenpipe — Main repository (desktop app + CLI)
- Documentation — Official docs
- API Reference — Full REST API reference (60+ endpoints)
- Scheduled Task Guide — Build custom pipes
- Discord — Community chat
- screenpipe/uniOCR — Native OCR for macOS, Windows, Linux
- screenpipe/audiopipe — Fast speech-to-text in Rust (Qwen3-ASR, CoreML, DirectML, CUDA)
- screenpipe-mcp — MCP server npm package; see setup instructions
Pipes are plugins that run inside Screenpipe and process your captured data.
- Obsidian Sync — Automatically sync screen activity and meeting notes to your Obsidian vault
- Notion Integration — Push captured data and summaries to Notion
- Meeting Notes — Auto-generate meeting summaries from audio transcriptions (Zoom, Google Meet, Teams, Webex)
- Pipe Store — Browse and install pipes from the in-app store
Screenpipe exposes an MCP (Model Context Protocol) server for retrieving recorded context and managing supported actions. Use Settings > Connections in the desktop app for the recommended setup: it configures the bundled runtime and SCREENPIPE_LOCAL_API_KEY. Manual setup requires a running Screenpipe instance; follow the official MCP installation guide. Retrieved context is available to the connected assistant and its configured model provider.
- Claude Desktop — Anthropic's desktop app
- Claude Code — CLI coding assistant
- Cursor — AI code editor
- VS Code + Cline — Autonomous coding agent for VS Code
- VS Code + Continue — Open-source AI code assistant
- Windsurf — AI IDE by Codeium
- Any MCP-compatible client
- mcpindex.net — MCP server directory
- pulsemcp.com — MCP server catalog
- topmcp.org — MCP server rankings
Screenpipe works with AI coding tools via MCP or direct API access:
- Cursor — Add Screenpipe MCP server in Cursor settings
- Claude Code — Configure as MCP server in
~/.claude.json - Cline — Add as MCP server in Cline extension settings
- Continue — Configure via Continue's MCP support
- OpenCode — MCP server integration
- Gemini CLI — Use via MCP or REST API
- Different AI / Note Companion — Obsidian plugin with Screenpipe integration for AI-powered meeting notes and screen activity search
- TanGentleman/screenpipe-python-client — Python client to debug and interact with Screenpipe's API
- baseballwalkerchris/screenpipe-hackathon — Hackathon project built on Screenpipe
- PrivateGPT — Screenpipe integration — Proposed integration for local screen/audio context
- Aider — Screenpipe for screen context — Proposed integration for coding session context
- Screenpipe Agentic Hackathon — LLM agent projects built on Screenpipe (Sprint.dev)
Screenpipe connects to external services via the Settings > Connections panel:
- Telegram — Send notifications and summaries via bot
- Slack — Post to channels via webhook
- Discord — Send messages via webhook
- Email — SMTP-based email notifications
- Todoist — Create tasks from screen activity
- Microsoft Teams — Post via webhook
- Salesforce
- HubSpot
- Pipedrive
- Building a Personal Knowledge Management System with Screenpipe — DEV.to tutorial
- Screenpipe Blog — Official blog with guides and updates
- Screenpipe Docs — Getting Started — Quick start guide
Contributions welcome! Include the project URL and a short description of how it integrates with Screenpipe.
If you've built something with Screenpipe, please open a PR to add it here.
