An MCP server for agent memory.
One endpoint, every assistant remembers.

Supermemory MCP is a managed memory server for any MCP-compatible client: Claude, ChatGPT, Cursor, your own agent runtime. Connect once and the assistant can search what it knew, save what it learns, and forget what expired, across sessions and across tools. OAuth sign-in, no API key, nothing to host.

Connect

Add one remote MCP server to any compatible client:

https://mcp.supermemory.ai/mcp

After you connect, your client opens supermemory in a browser. Sign in or create an account, then choose which spaces the client can access. It is OAuth, so there is no API key to paste and nothing to rotate.

Most clients that take a JSON configuration accept this shape:

{
  "mcpServers": {
    "supermemory": {
      "url": "https://mcp.supermemory.ai/mcp"
    }
  }
}

Setup and usage in the docs

The tools a client gets

You ask naturally; the assistant picks the right tool. No code, no tool names to learn.

Tool	What it does
search_memory	Semantic recall from one space. It returns matching memories; profile context comes from get_profile.
get_profile	Stable facts and recent context for a space: who the user is, their preferences, what they are working on.
add_memory	Save something worth keeping.
forget_memory	Forget what is outdated or should be removed.
list_documents	Browse stored source documents and their summaries.
get_document	Read the content of one stored document.
list_memories	Browse recent memories with their IDs and the documents they came from.
list_spaces	List the spaces the account can access, and resolve a space name to its key.
who_am_i	Show the signed-in account, its permissions and the active space.

Clients that support MCP Apps also open a space picker, a guided save form, a file upload and the memory graph right in the conversation.

Underneath is the same memory engine as the API: memories are extracted, linked to what the graph already knew, updated when facts change, and forgotten when they expire. A saved preference outlives the chat it was saved in.

Full tools reference

Spaces

A space keeps a project's documents, memories and profile context focused, so retrieval does not mix unrelated work.

Use separate spaces for an engineering launch, a research project, a legal matter, a personal knowledge base. Teammates collaborate inside the spaces they are allowed to read or write, and each client sees only the spaces you granted it. Name a space to use it for one request, or switch spaces to make it the default for what follows.

How spaces work

Clients

One memory layer, carried between the assistants you already use.

Claude Desktop and Claude Code.: Project context, decisions and preferences follow you between conversations and sessions. For Claude Code there is also a plugin with automatic recall and capture, built for coding work.
ChatGPT.: Add supermemory as a custom MCP server from the Plugins page, and let it recall your saved context inside a chat.
Cursor, and any other MCP client.: Anything that can connect to a remote MCP server can use the same endpoint, and most JSON configurations take the shape above.

Step-by-step setup for Claude, ChatGPT and Cursor, with a test that checks the apps share one memory, is in the MCP memory guide.

Claude Desktop guide

Local or managed

There are good local MCP memory servers, and for some setups they are the right call.

A local MCP memory server	Supermemory MCP
Runs on your machine. Fully private, fully yours to operate.	Managed for you: OAuth sign-in, no server to run, no database to back up.
One machine, one user, one client at a time.	The same memory in every client you connect, with team spaces and access control.
Stores what you tell it; retrieval is search.	Extracts, updates, derives and forgets; profiles are attached without a search.
You carry the upgrades and the breaking changes.	The engine improves underneath the same endpoint.

If your constraint is that nothing leaves the machine, this hosted endpoint isn't the fit. Self-hosted supermemory serves the API and the coding plugins but not this MCP server, and it sends telemetry unless you set SUPERMEMORY_DISABLE_TELEMETRY=1. If your constraint is that memory should work everywhere your team works, this endpoint is the shorter path.

Supermemory MCP in the docs

An MCP server for agent memory.One endpoint, every assistant remembers.