Integrations

Engram works with any MCP-compatible client. Here’s how to connect the most common ones.

Claude Desktop

Add Engram as a custom connector: in Claude, open Customize → Connectors, click +, choose Add custom connector, and enter https://mcp.getengram.app/mcp. Claude opens Engram’s sign-in page; approve the connection. The connector works in Claude on the web, desktop and mobile, and gives Claude Engram’s 8 core memory tools.

To connect with an API key instead, use Claude Desktop’s config file. It only runs local (stdio) servers, so Engram goes through mcp-remote (needs Node.js 18 or later):

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json


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

Restart Claude Desktop. With an API key, Claude gets the 8 core memory tools plus the vault and subscription tools. The Claude guide covers the details, including Windows.

Claude Code (CLI)

Add to your project’s .mcp.json:


{
  "mcpServers": {
    "engram": {
      "type": "http",
      "url": "https://mcp.getengram.app/mcp",
      "headers": {
        "Authorization": "Bearer engram_sk_live_your_key_here"
      }
    }
  }
}

Or add it to all your projects:


claude mcp add --transport http --scope user engram https://mcp.getengram.app/mcp \
  --header "Authorization: Bearer engram_sk_live_your_key_here"

Cursor

In Cursor Settings > MCP, add a new server:

Name: engram
Type: HTTP
URL: https://mcp.getengram.app/mcp
Headers: Authorization: Bearer engram_sk_live_your_key_here

Windsurf

Add to your Windsurf MCP configuration:


{
  "mcpServers": {
    "engram": {
      "serverUrl": "https://mcp.getengram.app/mcp",
      "headers": {
        "Authorization": "Bearer engram_sk_live_your_key_here"
      }
    }
  }
}

ChatGPT

ChatGPT connects over OAuth (not a static API key), so setup is different from the clients above:

Open the Plugin directory → search Engram and install it (or, in Developer mode, add a custom connector at https://mcp.getengram.app/mcp).
Sign in with your Engram account and authorize.

Capture in ChatGPT works two ways: save on request (“remember this”) and one-time bulk import of your ChatGPT data export. Unlike Claude Code or Cursor, ChatGPT has no background hook, so it can’t auto-record every message. See the full ChatGPT guide for details.

Custom MCP Client (TypeScript)

Use the official @modelcontextprotocol/sdk:


import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StreamableHTTPClientTransport } from "@modelcontextprotocol/sdk/client/streamableHttp.js";
 
const transport = new StreamableHTTPClientTransport(
  new URL("https://mcp.getengram.app/mcp"),
  {
    requestInit: {
      headers: {
        Authorization: "Bearer engram_sk_live_your_key_here",
      },
    },
  }
);
 
const client = new Client({ name: "my-agent", version: "1.0.0" });
await client.connect(transport);
 
// Create a conversation
const result = await client.callTool({
  name: "create_conversation",
  arguments: {
    title: "My first conversation",
    tags: ["test"],
  },
});
 
console.log(result);
// { content: [{ type: "text", text: '{"conversation_id":"conv_..."}' }] }

Custom MCP Client (Python)


from mcp import ClientSession
from mcp.client.streamable_http import streamablehttp_client
 
async def main():
    headers = {
        "Authorization": "Bearer engram_sk_live_your_key_here"
    }
 
    async with streamablehttp_client(
        "https://mcp.getengram.app/mcp",
        headers=headers
    ) as (read, write, _):
        async with ClientSession(read, write) as session:
            await session.initialize()
 
            result = await session.call_tool(
                "search",
                arguments={"query": "billing issue", "limit": 5}
            )
            print(result)

MCP Inspector

For testing and debugging, use the MCP Inspector:


npx @anthropic-ai/mcp-inspector

Enter your server URL (https://mcp.getengram.app/mcp) and API key. You can interactively call tools, inspect schemas, and see responses.

Using with AI Frameworks

LangChain / LangGraph

Engram works as an MCP tool provider. Use LangChain’s MCP integration to connect:


import { McpToolkit } from "@langchain/mcp";
 
const toolkit = new McpToolkit({
  servers: {
    engram: {
      url: "https://mcp.getengram.app/mcp",
      headers: {
        Authorization: "Bearer engram_sk_live_your_key_here",
      },
    },
  },
});
 
const tools = await toolkit.getTools();
// tools now contains create_conversation, append_messages, search, etc.

Vercel AI SDK

Use the MCP client adapter:


import { experimental_createMCPClient } from "ai";
 
const engram = await experimental_createMCPClient({
  transport: {
    type: "sse",
    url: "https://mcp.getengram.app/mcp",
    headers: {
      Authorization: "Bearer engram_sk_live_your_key_here",
    },
  },
});
 
const tools = await engram.tools();