Supermemory MCP
Long-term memory for agents: a remote MCP server, plugins for Claude Code, Cursor and Codex, and an API for your own apps
- 31,459 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Conversation content and extracted facts are stored in a third-party cloud
- Stored facts are injected into the agent's context automatically
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/supermemoryDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Conversation content and extracted facts are stored in a third-party cloud
- Stored facts are injected into the agent's context automatically
This is third-party code. Review the repository files before installing.
What it does
Supermemory extracts facts from conversations, maintains a user profile, handles updates and contradictions, and forgets stale information. The MCP server gives the agent three tools: memory saves or removes information, recall searches memories, and context injects the profile at the start of a conversation. Memory is split into projects to separate work, personal context or clients. Developers get an API and SDKs with hybrid search, connectors for Google Drive, Gmail, Notion, OneDrive and GitHub, and processing of PDFs, images and video, plus a version you can run locally.
Who it is for. For people who want an agent to remember preferences and project context across sessions.
Good fit when
- You are tired of re-explaining your stack and conventions to the agent
- You need shared memory across Claude Code, Cursor and Claude Desktop
- You are building a personalized product and do not want to run a vector database
Not a fit when
- Conversation content must not leave for an external cloud
- A CLAUDE.md or AGENTS.md file in the repo is enough
Example request
Remember that this project uses pnpm and avoids class componentsMCP
- Transport
- http
- Authentication
- OAuth
Limitations
The remote server stores memory in Supermemory's cloud and requires OAuth sign-in. The API needs a key and a paid plan beyond the free tier. The local version installs separately and works with a model of your choice, including Ollama. Cloud availability from Russia without a VPN is unverified.
How to disable. Remove the supermemory server from your MCP client config or uninstall the plugin, then delete stored data in the Supermemory dashboard if needed.
FAQ
Do I need an API key for MCP?
No, the remote MCP server uses OAuth: the client opens a browser to sign in and pick which spaces it can access.
Can memory be separated by project?
Yes, memory is scoped by projects defined with container tags.
Discussion
No questions or reviews yet.
Related
MCP serversAn MCP server for long-term agent memory: compresses dialogues into atomic facts, resolves pronouns and anchors relative time to absolute dates
iai-memory
iai-memory (Personal Memory Engine)
A local long-term memory MCP server for agents: keeps conversations verbatim and returns the right context at the start of a session
Local persistent memory for coding agents: hooks record every session, and Claude Code, Codex or Cursor recall past decisions through 54 MCP tools
Local memory for agents: stores conversations and project files verbatim and retrieves them with semantic search over MCP
Shared local memory for AI agents: Claude Code, Codex, Cursor and others find the agreements of past sessions in it and record new ones
A self-hosted MCP server for long-term agent memory: structured records, versioned rollback and a visual dashboard
README in short
The README addresses three groups: AI tool users, product developers, and people who want to self-host. The first get the MCP server URL, JSON config, tool list and links to separate plugin repos. The second get SDK install via npm or pip and examples of adding memories and fetching a profile. It also covers framework integrations and the local version.