JetBrains Integration Guide

Give JetBrains AI Assistant persistent memory using Engram. JetBrains IDEs connect through the standard mcp-remote bridge (the IDE speaks stdio MCP; the bridge handles Engram’s remote HTTP endpoint and auth).

Setup

1. Get an API key

2. Add the MCP server

Open Settings → Tools → AI Assistant → Model Context Protocol (MCP) → Add, choose the JSON option, and paste:


{
  "mcpServers": {
    "engram": {
      "command": "npx",
      "args": [
        "-y", "mcp-remote", "https://mcp.getengram.app/mcp",
        "--header", "Authorization: Bearer engram_sk_live_your_key_here"
      ]
    }
  }
}

(Requires Node.js on your PATH. Alternatively drop the two --header arguments to get a browser OAuth sign-in via mcp-remote instead of a key.)

3. Verify

In an AI Assistant chat, ask it to search Engram. If it can call the search tool, you’re connected.

Using memory in JetBrains

“Remember this fix” — stores the exchange verbatim
“Search memory for the schema discussion” — recalls exact prior context
The same memory follows you across every connected tool — IDE, terminal, ChatGPT, Claude.