Pieces

Pieces runs in the background on your computer and builds a searchable history of your work across apps. It's for people who need to recover a decision, find earlier research or resume a task after switching between tools. Memories stay on your machine by default.

It captures what you see in the focused application and what you copy and paste. Audio capture for meetings and conversations is opt-in. The chronological timeline brings together context from email, browser research, documents, chats and development tools, including Gmail, Slack, Notion, Figma and VS Code. You can search past activity or ask questions scoped to a time, source or topic.

That history also provides material for standup updates, meeting preparation, morning briefs and day recaps. Pieces can collect research on a topic into a summary or trace the work that led to a decision, so you don't have to reconstruct it across tabs and conversations.

Its MCP server makes remembered work available to Claude, Cursor, Codex, Antigravity and other MCP-compatible assistants. Memory storage is local by default; the MCP connection supplies work context to the assistant you use.

You control the capture. You can pause it, exclude particular apps or websites, and delete memories by time, source or capture method.

Capture, indexing and storage happen on-device. Current LLM features use cloud models and send the context selected for each request; the remainder of the stored history stays local.