screenpipe

screenpipe provides a context layer for AI workflows by capturing selected screen, audio, and application activity. Its public site describes locally stored history that users and approved agents can search to retrieve past work, meetings, and other context. A local API and MCP interface support integration with external assistants.

Potential applications include drafting meeting notes, recovering previous decisions, preparing work summaries, and building repeatable workflows. The distinction between local capture and downstream agent access matters: a connected model or automation can have its own data-handling behavior. Configure capture boundaries and review which tools may retrieve the history.