Persistent memory for AI agents — REST API, MCP, OAuth, CLI, dashboard. Semantic search, knowledge graph, automatic consolidation. One self-hosted service, every transport.
Milvus consolidation fully operational — embedding hydration end-to-end. ~1,838 tests.
End-to-end embedding hydration for Milvus deployments — consolidation no longer produces 0 clusters/associations. consolidator._get_memories_for_horizon now passes include_embeddings=True to bulk-read methods, ensuring clustering receives real vectors. Completes a 4-PR series (#872, #878, #881, #885). (PR #885, @henry201605)
Upgraded GitPython to the latest secure release, resolving 3 high-severity CVEs in the transitive wandb dependency: path traversal (arbitrary file write/delete), newline injection enabling RCE via core.hooksPath, and a bypass of the prior CVE patch. (PR #886)
Opt-in two-tier fallback in retrieve_memories (fallback=True) for sparse semantic results: BM25 exact-match, then tag-intersection, merged and de-duplicated up to n_results. Default off — existing callers unaffected. (PR #883, @filhocf)
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.