Pieces

Paragraph 1: Pieces functions as a memory layer that captures work context and makes it searchable, enabling users and AI tools to access remembered decisions, sources, and discussions. It runs in the background on a computer and builds a memory from activities across apps and websites to help users understand, do, or create next. It targets individuals and teams who rely on distributed tools and need grounded context for AI-assisted work.

Paragraph 2: The system remembers activities chronologically—research, chats, emails, notes, meetings—and allows searching by time, topic, person, or tool. It provides prompts scoped by time, source, or topic to locate or curate precise information. It also generates standup updates and briefs from real context, helping users resume work efficiently.

Paragraph 3: Core capabilities include capturing memories from applications and websites, providing one-click recall, and enabling cross-tool context transfer into MCP-ready AI tools. It supports persistent memory across tools and offers control over what is captured and when, with enterprise options for policy and governance.

Paragraph 4: Pieces is available with plans including Pro and Enterprise, with on-device memory by default and governance options for teams. It integrates with MCP and supports provider flexibility, including the option to bring your own API keys. It emphasizes on-device storage, organization-level controls, and enterprise rollout features.

Use cases

• Find a source article across many tabs and references
• Compile a topic into a single summary from multiple sources
• Trace the origin of a decision across discussions
• Resume a deep dive where it left off