Pieces is a local-first productivity platform that captures context from your workflow and turns it into useful memory for you and your AI tools.
With PiecesOS, the desktop app, and the Pieces MCP server, Pieces can remember notes, browser activity, conversations, documents, code, and other workflow context so it is easier to recover past work, answer questions, and stay in flow without shipping your context off-device by default.
main-app.mp4
Pieces helps you work with memory instead of hunting for context.
- Long-Term Memory captures and organizes workflow context over time.
- Timeline gives you a central place to revisit activity, summaries, and past work.
- Conversational workflows let you ask questions, generate content, and retrieve useful context from what you have already done.
- Pieces MCP brings that context into MCP-compatible AI clients.
- PiecesOS is the local foundation that powers capture, memory, and context.
- The Desktop App is your main hub for browsing memories, managing context, and interacting with Pieces directly.
- Pieces MCP connects that local context to AI clients like Cursor, Claude Desktop, GitHub Copilot, and Codex CLI.
- Recover work you have already done instead of reconstructing it from memory.
- Search across your real workflow context, not just filenames or isolated snippets.
- Generate summaries, answers, and updates from actual activity.
- Keep control of your data with local-first privacy and security.
- Install the Pieces Desktop App or install PiecesOS manually.
- Turn on Long-Term Memory so Pieces can start capturing context.
- Set up Pieces MCP for your preferred AI client.
- Use the prompting guide to get better results from your personal context.
- Cursor
- GitHub Copilot
- Claude Desktop
- VS Code
- JetBrains IDEs
- OpenAI Codex CLI
- Get started with Pieces MCP
This repository powers the public GitHub profile for the pieces-app organization. The content in profile/README.md is mirrored in the top-level README.md so the repository page and org profile stay in sync.