Claude-Mem
Persistent context infrastructure that captures agent activity and makes summaries available in later sessions. It includes adapters for multiple agent hosts, with local and hosted memory paths.
Utility: Agent operations; kind: infrastructure.
Stars: 107k+; Starred in September: 2k+.
Metrics updated: 31 Oct 2026. Profile reviewed: 31 Oct 2026.
Utilities
- Retain compressed observations across coding-agent sessions.
- Retrieve earlier context when resuming work on a project.
Who it is for
- Agent users and developers who need inspectable cross-session memory.
Requirements
- The README lists Node.js 20+, Bun, uv, and SQLite for the relevant local setup.
- Install the adapter for the selected host and inspect its memory-provider configuration.
Limitations
- Compressed observations can omit details or preserve incorrect conclusions.
- Provider defaults vary by adapter; the reviewed Grok Bot setup defaults to hosted CMEM Pro, not local storage.
Cost and privacy notes
- Hosted memory, summarization models, and local infrastructure can have different costs.
- Captured tool observations may include secrets or personal data; review exclusions, storage, retention, and provider defaults.
Illustrative example (untested)
Illustrative, not tested: retain a non-sensitive bug-fix observation and verify what a later session retrieves.
License
Apache-2.0 for the documented public core and adapters; hosted/cloud, team, and enterprise areas are outside that grant.