MCP server
Memory MCP server
The MCP project's Memory reference server gives AI a local knowledge graph to store people, things and facts and recall them later.
- Official server
- Knowledge and memory
At a glance
- Maintained by
- Model Context Protocol project (reference server)
- Tools exposed
- 9 listed below
- Reviewed
- September 24, 2026
Why it matters
What it does and why it matters
By default, an assistant forgets everything when a conversation ends. The Memory server gives it a place to remember: a knowledge graph stored in a local file, made of entities such as people, companies and projects, relations between them and observations about them. The assistant can add, search, read and delete this information.
For a single user or a small team, it is a simple way to let an assistant remember preferences, recurring context and facts it should not have to ask for again.
It is a reference server from the Model Context Protocol project, and it stores data in a local file with no outside service. Decide what the assistant may remember, avoid storing sensitive personal data and review the graph from time to time. For company-wide knowledge, a proper knowledge base with permissions is a better fit.
Because the memory is a plain file, it is easy to back up, inspect and delete, which makes it a good way to learn how assistant memory works before building something larger.
Use cases
Best business use cases
Personal preferences
Client context
Relationships
Recall
Learning memory
Tools
Available tools
Tool names as published by the maintainer at review time. Versions change, so check the repository for the current list.
Read and write: Read and write: stores and edits a local knowledge graph file; no outside services.
-
create_entitiesAdd new entities (people, things, ideas) to the knowledge graph.
-
create_relationsLink two entities with a named relationship.
-
add_observationsAdd new facts to an existing entity.
-
delete_entitiesRemove entities and their relations.
-
delete_observationsRemove specific facts from an entity.
-
delete_relationsRemove relationships between entities.
-
read_graphRead the whole knowledge graph.
-
search_nodesSearch entities by name, type or fact text.
-
open_nodesFetch specific entities by name.
Workflows
Example workflows
Remember a new client
-
1
Tell the assistant about a new client.
-
2
It creates entities for the client and contacts.
-
3
It adds relations to your team.
-
4
It stores key preferences.
-
5
Later chats recall them.
Review stored memory
-
1
Ask to read the whole graph.
-
2
The assistant lists entities and facts.
-
3
You spot outdated items.
-
4
It deletes them on request.
-
5
The memory stays accurate.
Prompts
Example prompts
“Remember that Harbor Group prefers invoices at the start of each month.”
“What do you know about the Acme project?”
“Show everything you have stored and delete anything about the old office.”
Setup
Installation and configuration
Examples for Claude Desktop and Claude Code. Replace the placeholder values with your own keys, and store secrets outside version control.
Claude Desktop: claude_desktop_config.json
{
"mcpServers": {
"memory": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-memory"
]
}
}
}Claude Code: terminal
claude mcp add memory -- npx -y @modelcontextprotocol/server-memory- Local server: npm: @modelcontextprotocol/server-memory (stdio).
- Needs: None. Optional MEMORY_FILE_PATH env var sets where the JSONL graph file is stored..
Security
Security considerations
-
What gets stored
Decide what may be remembered and avoid sensitive personal data.
-
Local file
The graph is a local file. Protect and back it up.
-
Review regularly
Check the graph for outdated or wrong facts.
-
Reference server
Treat it as a reference implementation.
Related servers
Servers that pair well with it
Qdrant
Qdrant's official MCP server gives AI a semantic memory: store information in a Qdrant collection and find it again by meaning.
Obsidian
A community MCP server that lets AI list, read, search, edit, append to and delete notes in an Obsidian vault through the Local REST API plugin.
Sequential Thinking
The MCP project's Sequential Thinking reference server gives AI a structured way to break a problem into steps and revise them.
Keep exploring
Solutions and services that use it
Related services
View all related services- MCP integration Connect AI assistants to your CRM, files, databases and tools through Model Context Protocol servers, with safe permissions.
- RAG knowledge assistants Assistants that answer questions from your own documents and show where each answer came from.
- AI Product Development AI agents, knowledge assistants, copilots and document automation built into the way your team already works.
Solutions
View all solutions- Knowledge assistant (RAG) Ask a question in plain words and get an answer from your own documents, with links to the sources.
- Internal help desk assistant An assistant in Slack or Teams that answers policy and how-to questions from your handbooks and opens tickets when needed.
- AI sales research agent Short, sourced briefings on each prospect before a call, drafted by an AI agent from public information and your CRM.
- AI support agent An assistant that answers routine questions from your own content, checks orders and hands anything else to a person.
Industries
View all industries- Healthcare Patient booking, intake forms, internal knowledge assistants and admin automation for clinics and care providers.
- Legal Contract review assistants, knowledge search, intake and document automation for law firms and in-house legal teams.
- Professional services Client portals, proposal drafting, knowledge assistants and internal tools for consultancies, agencies and firms.
Guides and articles
View all guides and articles- What is RAG and when do you need it? How retrieval-augmented generation lets AI answer from your documents, when it fits and how to build it well.
- RAG vs fine-tuning Two ways to make AI work with your knowledge, compared by cost, accuracy and upkeep.
- What are AI agents and how do businesses use them? What AI agents are, real business uses, the controls they need and how to start with one safely.
- How to build an AI workflow with MCP servers Plan, connect and test an AI workflow that uses MCP servers, from a single task to a reliable process.
- AI chatbot vs AI agent What separates a chatbot that answers from an agent that acts, and when you need each.
AI models
View all ai models- Gemini Embedding 2 Google's current embedding model, multimodal: it embeds text, images, video, audio and PDFs for search and retrieval.
- text-embedding-3-large OpenAI's most capable embedding model for search and retrieval, with 3,072-dimension vectors and support for English and other languages.
- text-embedding-3-small OpenAI's efficient, lowest-cost embedding model, with 1,536-dimension vectors for search, retrieval and similarity at scale.
- Cohere Embed v4 Cohere's multimodal embedding model for enterprise search, embedding text, images and mixed documents, with flexible vector sizes and long inputs.
- Cohere Rerank 4 Cohere's multilingual rerank models, which sort search results by relevance, in a best-quality Pro version and a low-latency Fast version.
- Voyage 4 Voyage AI's current embedding series, now part of MongoDB, with large, standard and lite models that share one embedding space.
Glossary terms
View all glossary terms- RAG RAG, or retrieval-augmented generation, is a method where an AI system first finds relevant passages in your documents, then answers using only those passages.
- AI agent An AI agent is software that uses a language model to plan steps, call tools and act toward a goal, instead of only answering one question.
- Chatbot A chatbot is software that holds a conversation with people through text or voice, answering questions or completing simple tasks.
- Embeddings Embeddings are lists of numbers that represent the meaning of text or images, so software can find items that are similar in meaning.
- Semantic search Semantic search finds results by meaning rather than exact words, so a search for "reset my login" can find an article titled "Forgot your password".
- Vector database A vector database stores embeddings, lists of numbers that represent meaning, and quickly finds the ones most similar to a query.
FAQ
Questions people ask us
Have a question that is not here? Ask us directly.
In a local file, as a knowledge graph.
It is best for individuals or small teams. Company knowledge needs permissions and a proper knowledge base.
It is a reference server from the MCP project.