Shared memory for Cursor

When you add Glen as an MCP server in Cursor, your agent gains a single tool that does two things in one round trip: it pulls the context relevant to what you are working on, and it records new facts worth remembering. The first time someone's Cursor learns that your auth lives behind a particular middleware, or that a service must never be called from the edge runtime, that observation lands in the organization's shared memory. The next time anyone on the team opens Cursor and touches related code, that knowledge surfaces automatically, no re-explaining, no rediscovering.

Because Glen is org-scoped rather than per-user, the value compounds with team size. A new hire's Cursor inherits the accumulated memory of everyone who came before them, so onboarding stops meaning the agent relearns the codebase from scratch. And because Cursor supports both global and project-scoped MCP config, you can wire Glen in once for yourself or ship it in a repo so every contributor's Cursor connects automatically. Setup is a single mcpServers entry; from there the memory grows on its own as your team works.

Add Glen to Cursor

{
  "mcpServers": {
    "glen": {
      "url": "https://tryglen.com/api/mcp"
    }
  }
}

FAQ

Where do I configure Glen in Cursor?

Add it to the mcpServers object in ~/.cursor/mcp.json for global use, or in a project's .cursor/mcp.json so every contributor on that repo connects automatically.

Does each teammate get separate memory?

No. Glen memory is scoped to the organization, so every teammate's Cursor reads and writes the same shared store and inherits what others have learned.