Back to Tools

screenpipe is a local-first work-memory tool for individuals, teams, and AI agents. On macOS, Windows, and Linux, it continuously captures accessibility text, screen content, and audio transcripts into a searchable history stored on the device by default. A local REST API, MCP server, CLI, and SQLite access can supply real work context to tools such as Claude and Codex, while trigger-based workflows can create meeting notes, follow-ups, and SOPs. Cloud AI, sync, and connectors are opt-in; the code is source-available under the Screenpipe Commercial License.

screenpipe homepage screenshot

Features

  • Local-first screen, app, and audio history
  • Accessibility text, OCR, and audio transcription
  • Natural-language search across prior work context
  • REST API, MCP, CLI, and SQLite access
  • Trigger-based meeting notes, follow-ups, and SOPs
  • App exclusions, privacy filters, and sensitive-data controls

Use Cases

  • Recall information previously viewed or handled
  • Create meeting notes and action items
  • Give AI assistants real work context
  • Turn repeated work into SOPs and automations
  • Find decisions, files, and communications across apps
  • Build a local personal knowledge and work-memory archive

FAQ

Recording and indexing stay on the device by default. Context may leave the device only when you explicitly enable cloud AI, sync, export, or third-party connectors.

No. Its code is viewable and self-hostable, but it uses the Screenpipe Commercial License, making it source-available rather than MIT-licensed or OSI-approved open source.

Alternatives and related tools