SkillFoxx

SimpleMem MCP

An MCP server for long-term agent memory: compresses dialogues into atomic facts, resolves pronouns and anchors relative time to absolute dates

  • 3,875 ★ on the repository
  • Last commit
  • MIT
MCP serverOfficial

Extended access

Why this level

  • The cloud option stores the user's dialogue content on a third-party service
  • It needs an OpenRouter token that spends budget on fact extraction and embeddings
All reasons and checks

aiming-lab/simplememAuthor:aiming-lab

Install

In your terminal, with SkillFoxx CLI

npx skillfoxx add mcp/simplemem

Detects the agents on your machine, shows the access level and pins the version.

Other ways to install

Access and checks

How we review

Extended accessThe tool runs code or makes network calls and only reads external data.

  • The cloud option stores the user's dialogue content on a third-party service
  • It needs an OpenRouter token that spends budget on fact extraction and embeddings

This is third-party code. Review the repository files before installing.

What it does

SimpleMem turns a dialogue history into compact, self-contained facts: it replaces pronouns like 'he' and relative times like 'tomorrow' with concrete names and dates instead of storing raw conversation text. For queries it uses hybrid retrieval combining semantic search, keyword matching and metadata filters, and can decompose complex questions into sub-queries and refine the answer. Each user's data lives in a separate table with token authorization; the connection is MCP over Streamable HTTP, either to the cloud service or to a self-hosted Python server.

Who it is for. For anyone who wants their AI assistant to remember things across sessions without manually re-pasting context each time.

Good fit when

  • You need the assistant to remember facts and decisions from past conversations
  • You need to search accumulated history by meaning, not just keywords
  • You work across several MCP clients and want a shared memory store between them

Not a fit when

  • Dialogue data cannot go to a third-party cloud service under your privacy policy
  • In-session memory without persistence across runs is already enough

Example request

Remember via SimpleMem that I prefer TypeScript and I'm working on project X, and pull it back from memory next session

MCP

Transport
http
Authentication
key is optional

Environment variables

JWT_SECRET_KEY secret
Secret for signing tokens when self-hosting.
ENCRYPTION_KEY secret
32-byte encryption key for stored API keys when self-hosting.

Limitations

The cloud option requires an OpenRouter key and an auth token; fact extraction and embeddings are computed via an LLM and embedding model on OpenRouter, so usage is paid. Self-hosting removes the cloud dependency but needs a Python environment and your own infrastructure.

How to disable. Remove the simplemem block from the MCP client configuration. For self-hosting, stop the python run.py process.

FAQ

Can it be self-hosted?

Yes, the MCP folder in the repo contains a self-hosted Python server started with python run.py.

What tools are available to the agent?

memory_add, memory_add_batch, memory_query, memory_retrieve, memory_stats and memory_clear cover the full memory lifecycle.

How is SimpleMem different from storing raw conversation logs?

The server turns dialogue history into compact self-contained facts: it replaces pronouns and words like tomorrow with concrete names and absolute dates. That keeps memory understandable outside the original conversation.

How do I connect SimpleMem to Claude Desktop and Cursor?

Add the simplemem server to the MCP config with the address from the install block. The hosted server works in Claude Desktop, Cursor and Claude Code, and a self-hosted option exists. Dialogue data goes to the server, keep that in mind.

Discussion

No questions or reviews yet.

MCP servers

A self-hosted MCP server for long-term agent memory: structured records, versioned rollback and a visual dashboard

MCP serverExtended access1.4K
Official

Long-term memory for agents: a remote MCP server, plugins for Claude Code, Cursor and Codex, and an API for your own apps

MCP serverExtended access31.6K
Official

The Memanto memory MCP server: an agent stores facts and decisions and finds them in later sessions, even as of a date.

MCP serverExtended access2.4K

A long-term memory MCP server for AI agents: facts fade along a forgetting curve and related records are compressed into a summary

MCP serverExtended access272
Official

A long-term memory system for AI agents: the agent stores experience and facts, then retrieves them through a built-in MCP server

MCP serverBroad access60.7K

README in short

The MCP server README describes SimpleMem as a production long-term memory service with lossless semantic compression, pronoun resolution, time anchoring and hybrid retrieval. It shows a quick start via the mcp.simplemem.cloud cloud service and a self-hosted Python option. A table of MCP tools and a sample client JSON config with a url field and an authorization header are included.

README badge

Report a problem