mcp-memory-service v11.12 — Now Available
Memory for AI Agents
Persistent memory for AI agents — REST API, MCP, OAuth, CLI, dashboard. Semantic search, knowledge graph, automatic consolidation. One self-hosted service, every transport.
mcp-memory-service v11.12 — Now Available
Persistent memory for AI agents — REST API, MCP, OAuth, CLI, dashboard. Semantic search, knowledge graph, automatic consolidation. One self-hosted service, every transport.
Interactive 3D knowledge graph — every memory a glowing node, every relationship a curved edge.
MINOR release: the first since development moved back to GitHub. 73 merged pull requests, 43 of them from outside the maintainer, with the weight of them on retrieval correctness and consolidation. ~3,290 tests.
Four separate paths asked for the k nearest embeddings first and filtered afterwards, so excluded rows ate the candidate budget and a search could come back empty while matching memories existed. Eligibility now scopes the scan on sqlite-vec, Cloudflare over-fetches before applying a tag filter, match_all means AND, and compact search applies tags before the limit. (#1128, #1218, #1197, #1189)
The duplicate check is symmetric and ran once per memory against the whole list, so both members of a near-duplicate pair were marked deletable, and that path skips the time-horizon guard. Forgetting also could never reach the stale tail, and the health monitor reported fixed strings instead of runtime state. (#1180, #1167, #1166)
The ONNX backend honors MCP_EMBEDDING_MODEL instead of treating it as a label, so a multilingual model runs without torch. MCP_HTTP_ROOT_PATH now reaches the ASGI root path, OpenAPI, the OAuth issuer, static mounts and the browser-side links, so a deployment behind a stripped prefix serves the dashboard and API through it. (#1242, #1198)
Claude.ai connects to your memory server through Streamable HTTP with OAuth 2.1 authentication.
Battle-tested with comprehensive testing and optimized for performance.
Talks and coverage from the community.
Heavybit's Humans in the Loop deep dive talks to maintainer Heinrich Krupp about agent amnesia and why persistent memory is the missing infrastructure layer for agentic systems.
Maintainer Heinrich Krupp presents mcp-memory-service at the AI Tinkerers Zürich meetup — persistent memory architecture, semantic search, and multi-agent memory sharing. Watch the full video →
Install from PyPI and connect your AI agents to persistent memory.
I help teams run self-hosted memory in production — OAuth, hybrid sync, multi-client rollout. Free 30-minute intro call.
Get in touch