Persistent memory for AI agents — REST API, MCP, OAuth, CLI, dashboard. Semantic search, knowledge graph, automatic consolidation. One self-hosted service, every transport.
CLI lifecycle commands + lazy imports for <3s startup. Security fixes. ~1,803 tests.
New memory launch|stop|restart|info|health|logs commands manage the background HTTP server with cross-platform PID tracking and health polling. restart threads --storage-backend and --debug flags from the running process automatically. (PR #841, @creativelaides)
Heavy ML dependencies (torch, transformers, sentence-transformers) are now lazy-loaded โ imported only when actually invoked. memory --help and lifecycle commands start instantly without waiting for model initialization. (PR #841, @creativelaides)
Command injection in launch patched (safe arg list replaces -c string interpolation); file-descriptor leak after Popen closed; MCP_ALLOW_ANONYMOUS_ACCESS pass-through restored; PID-reuse detection via create_time + cmdline_hint. (PR #841)
Claude.ai connects to your memory server through Streamable HTTP with OAuth 2.1 authentication.
Battle-tested with comprehensive testing and optimized for performance.
Install from PyPI and connect your AI agents to persistent memory.