[{"data":1,"prerenderedAt":4},["ShallowReactive",2],{"readme:agent-memory":3},"\u003Ch1>Neo4j Agent Memory\u003C\u002Fh1>\n\u003Cp>A graph-native memory system for AI agents. Store conversations, build knowledge graphs, and let your agents learn from their own reasoning -- all backed by Neo4j.\u003C\u002Fp>\n\u003Cp>\u003Ca href=\"https:\u002F\u002Fneo4j.com\u002Flabs\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FNeo4j-Labs-6366F1?logo=neo4j\" alt=\"Neo4j Labs\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fneo4j.com\u002Flabs\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FStatus-Experimental-F59E0B\" alt=\"Status: Experimental\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fcommunity.neo4j.com\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FSupport-Community-6B7280\" alt=\"Community Supported\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneo4j-labs\u002Fagent-memory\u002Factions\u002Fworkflows\u002Fci-python.yml\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fgithub.com\u002Fneo4j-labs\u002Fagent-memory\u002Factions\u002Fworkflows\u002Fci-python.yml\u002Fbadge.svg\" alt=\"Python CI\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneo4j-labs\u002Fagent-memory\u002Factions\u002Fworkflows\u002Fci-typescript.yml\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fgithub.com\u002Fneo4j-labs\u002Fagent-memory\u002Factions\u002Fworkflows\u002Fci-typescript.yml\u002Fbadge.svg\" alt=\"TypeScript CI\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fbadge.fury.io\u002Fpy\u002Fneo4j-agent-memory\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fbadge.fury.io\u002Fpy\u002Fneo4j-agent-memory.svg\" alt=\"PyPI version\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002F@neo4j-labs\u002Fagent-memory\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fnpm\u002Fv\u002F@neo4j-labs\u002Fagent-memory.svg\" alt=\"npm version\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Fneo4j-agent-memory\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fpyversions\u002Fneo4j-agent-memory.svg\" alt=\"Python versions\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fopensource.org\u002Flicenses\u002FApache-2.0\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-Apache%202.0-blue.svg\" alt=\"License\" \u002F>\u003C\u002Fa>\u003C\u002Fp>\n\u003Ch2>What It Does\u003C\u002Fh2>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fneo4j-labs\u002Fagent-memory\u002FHEAD\u002Fimg\u002Fmemory-graph-model.png\" alt=\"The Neo4j Agent Memory data model\" \u002F>\u003C\u002Fp>\n\u003Ctable>\n\u003Cthead>\n\u003Ctr>\n\u003Cth>Short-Term Memory\u003C\u002Fth>\n\u003Cth>Long-Term Memory\u003C\u002Fth>\n\u003Cth>Reasoning Memory\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003C\u002Fthead>\n\u003Ctbody>\u003Ctr>\n\u003Ctd>Conversations &amp; messages\u003C\u002Ftd>\n\u003Ctd>Entities, preferences, facts\u003C\u002Ftd>\n\u003Ctd>Reasoning traces &amp; tool usage\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>Per-session history\u003C\u002Ftd>\n\u003Ctd>Knowledge graph (\u003Ca href=\"https:\u002F\u002Fneo4j.com\u002Flabs\u002Fagent-memory\u002Fexplanation\u002Fpoleo-model\" rel=\"nofollow ugc noopener\">POLE+O model\u003C\u002Fa>)\u003C\u002Ftd>\n\u003Ctd>Learn from past decisions\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>Vector + text search\u003C\u002Ftd>\n\u003Ctd>Entity resolution &amp; dedup\u003C\u002Ftd>\n\u003Ctd>Similar task retrieval\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftbody>\u003C\u002Ftable>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fneo4j-labs\u002Fagent-memory\u002FHEAD\u002Fimg\u002Fextraction-pipeline.png\" alt=\"The Neo4j Agent Memory entity extraction pipeline\" \u002F>\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Plus:\u003C\u002Fstrong> multi-stage entity extraction (spaCy \u002F GLiNER \u002F LLM), relationship extraction (GLiREL), background enrichment (Wikipedia \u002F Diffbot), geospatial queries, \u003Ca href=\"#mcp-server\" rel=\"nofollow ugc noopener\">MCP server\u003C\u002Fa> with 16 tools, and integrations with \u003Ca href=\"#framework-integrations\" rel=\"nofollow ugc noopener\">LangChain, Pydantic AI, Google ADK, Strands, CrewAI, and more\u003C\u002Fa>.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Production features:\u003C\u002Fstrong> adopt an existing Neo4j graph as long-term memory (\u003Ccode>client.schema.adopt_existing_graph(...)\u003C\u002Fcode>), multi-tenant scoping (\u003Ccode>user_identifier=\u003C\u002Fcode>), fire-and-forget \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneo4j-labs\u002Fagent-memory\u002Fblob\u002FHEAD\u002Fexamples\u002Fbuffered-writes\u002F\" rel=\"nofollow ugc noopener\">buffered writes\u003C\u002Fa> (\u003Ccode>client.buffered.submit(...)\u003C\u002Fcode>), \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneo4j-labs\u002Fagent-memory\u002Fblob\u002FHEAD\u002Fexamples\u002Faudit-trail\u002F\" rel=\"nofollow ugc noopener\">consolidation primitives\u003C\u002Fa> (\u003Ccode>client.consolidation.dedupe_entities(...)\u003C\u002Fcode>), an \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneo4j-labs\u002Fagent-memory\u002Fblob\u002FHEAD\u002Fexamples\u002Feval-harness\u002F\" rel=\"nofollow ugc noopener\">eval harness\u003C\u002Fa> (\u003Ccode>client.eval.run(suite)\u003C\u002Fcode>), and explicit \u003Ccode>:TOUCHED\u003C\u002Fcode> audit edges from reasoning steps to entities.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Bring your own model:\u003C\u002Fstrong> \u003Ccode>MemorySettings.embedding\u003C\u002Fcode> and \u003Ccode>MemorySettings.llm\u003C\u002Fcode> accept a provider-string shorthand (\u003Ccode>\"anthropic\u002Fclaude-3-5-sonnet-latest\"\u003C\u002Fcode>, \u003Ccode>\"BAAI\u002Fbge-small-en-v1.5\"\u003C\u002Fcode>) or a Provider instance. Native adapters for OpenAI, Anthropic, Bedrock, Vertex AI, and sentence-transformers; LiteLLM universal fallback covers 100+ providers (Cohere, Voyage, Groq, Together, Mistral, Ollama, ...). See the \u003Ca href=\"https:\u002F\u002Fneo4j.com\u002Flabs\u002Fagent-memory\u002Fhow-to\u002Fmigrate-to-providers.html\" rel=\"nofollow ugc noopener\">provider migration guide\u003C\u002Fa>. \u003Cem>(These configure the \u003Cstrong>self-hosted\u003C\u002Fstrong> backend; on NAMS, embedding and extraction run server-side.)\u003C\u002Fem>\u003C\u002Fp>\n\u003Ch2>SDKs\u003C\u002Fh2>\n\u003Cp>\u003Ccode>neo4j-labs\u002Fagent-memory\u003C\u002Fcode> ships two SDKs with the same memory model, both\nbacked by the \u003Ca href=\"https:\u002F\u002Fneo4j.com\u002Flabs\u002Fagent-memory\u002Freference\u002Frest-api\" rel=\"nofollow ugc noopener\">NAMS\u003C\u002Fa>\nhosted service. Pick the one that matches your stack — mixed Python +\nTypeScript agents read and write the same memory.\u003C\u002Fp>\n\u003Ctable>\n\u003Cthead>\n\u003Ctr>\n\u003Cth>Language\u003C\u002Fth>\n\u003Cth>Package\u003C\u002Fth>\n\u003Cth>Install\u003C\u002Fth>\n\u003Cth>Docs\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003C\u002Fthead>\n\u003Ctbody>\u003Ctr>\n\u003Ctd>Python\u003C\u002Ftd>\n\u003Ctd>\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Fneo4j-agent-memory\u002F\" rel=\"nofollow ugc noopener\">\u003Ccode>neo4j-agent-memory\u003C\u002Fcode>\u003C\u002Fa>\u003C\u002Ftd>\n\u003Ctd>\u003Ccode>pip install neo4j-agent-memory\u003C\u002Fcode>\u003C\u002Ftd>\n\u003Ctd>\u003Ca href=\"https:\u002F\u002Fneo4j.com\u002Flabs\u002Fagent-memory\u002Fsdks\u002Fpython\" rel=\"nofollow ugc noopener\">Python SDK docs\u003C\u002Fa>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>TypeScript\u003C\u002Ftd>\n\u003Ctd>\u003Ca href=\"https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002F@neo4j-labs\u002Fagent-memory\" rel=\"nofollow ugc noopener\">\u003Ccode>@neo4j-labs\u002Fagent-memory\u003C\u002Fcode>\u003C\u002Fa>\u003C\u002Ftd>\n\u003Ctd>`npm install @neo4j-labs\u002Fagent-memor\u003C\u002Ftd>\n\u003Ctd>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftbody>\u003C\u002Ftable>\n",1787445486841]