screenpipe/screenpipe

YC (S26) | Open Computer History | Continuously record your company computer work, map your workflows, help you find work worth automating, and power your agents' context

What it solves

Screenpipe addresses the challenge of fragmented digital workflows by creating a searchable, AI-powered memory of your computer activity. It helps users recall past tasks, automate repetitive work, and provide AI agents with the necessary context to assist with daily operations.

How it works

Screenpipe runs locally on your machine, continuously capturing screen frames, system audio, microphone input, and keyboard activity. It uses event-driven recording—triggering captures based on app switches or user activity rather than constant video recording—to keep resource usage efficient. This data is stored in a local SQLite database, which can be queried via a REST API, an MCP (Model Context Protocol) server for AI assistants, or through "Pipes"—customizable AI agents defined in markdown files that can automate tasks based on your history.

Who it’s for

  • Knowledge workers who need to track their time or recall past meetings and documents.
  • Developers who want to provide AI coding assistants (like Cursor or Claude) with context about their active projects.
  • Teams and enterprises looking to deploy AI workflows with deterministic, OS-level control over data privacy and access permissions.

Highlights

  • Local-First Architecture: Captures and stores data on your device by default, with optional cloud features for sync and team management.
  • Event-Driven Capture: Records only meaningful changes (app switches, clicks, typing) to minimize CPU and storage impact.
  • Plugin System (Pipes): Allows you to define scheduled AI agents using simple markdown files to automate tasks like meeting summaries or time tracking.
  • Broad AI Integration: Works with popular AI coding assistants via the Model Context Protocol (MCP) and supports local inference models like Ollama.
  • Deterministic Permissions: Provides granular, YAML-based access controls for AI agents, ensuring they only access authorized apps, windows, or data types.

Related

  • Project
    GeminiLight/MindOS

    MindOS is a local-first knowledge base that gives you and your AI agents a shared, persistent memory. It stores your knowledge as plain Markdown files, connects agents via MCP and Skills, and helps you turn reviewed conversations into reusable SOPs so agents start from a better baseline.

  • Project
    caspianmoon/memoripy

    An evidence-first, local memory runtime for AI agents that prevents noise ingestion and provides verifiable, bitemporal records with retrieval receipts.

  • Project
    FutureUniant/WorkShadow

    A local-first desktop work log application that uses AI for semantic search, automated report generation, and Q&A over personal work history while keeping data on the user's machine.

  • Project
    LYiHub/pub-local-jarvis

    A local multimodal desktop assistant for Windows that perceives screen and system audio to provide real-time gaming companionship, course notes, and contextual AI interaction.

  • Project
    EXXETA/exxperts

    A local-first AI agent framework that uses approval-gated memory to create persistent AI colleagues that learn and remember only what the user explicitly approves.