Granola

MCP server for semantic search across Granola meeting notes with speaker attribution and theme extraction.

javascript

0

GitHub Stars

javascript

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "mishkinf-granola-mcp": {
      "command": "node",
      "args": [
        "/path/to/granola-mcp/dist/mcp/server.js"
      ],
      "env": {
        "OPENAI_API_KEY": "sk-...",
        "GRANOLA_DATA_DIR": "/path/to/granola-mcp/export"
      }
    }
  }
}

This MCP server provides semantic search over Granola meeting notes, extracting insights, themes, and quotes with speaker attribution, and exposing these capabilities to MCP-enabled clients for fast, local searches and structured results.

How to use

You connect an MCP client to this server as a local or remote service to perform semantic searches across your Granola transcripts. The server reads your Granola export data, builds a vector index, and returns summaries plus relevant quotes in response to your search queries. Use it to quickly surface pain points, feature requests, decisions, and speaker-attributed quotes from your meetings. The client can issue search, search by theme, or request document lists and full document details via defined tools.

How to install

Prerequisites you need to prepare before installing: Node.js 18 or newer, Granola desktop app installed and logged in, and an OpenAI API key for embeddings and insight extraction.

Install and build the server from source, then run the build to generate the MCP-ready runtime.

npm install
npm run build

Configuration and running note

Once built, you run the MCP server locally as a stdio server. You will launch it with Node and point to the compiled server file, providing your Granola export data directory and your OpenAI API key as environment variables.

A representative runtime configuration shown below demonstrates how to start the MCP server using Node and a local build path, with the required environment variables.

{
  "mcpServers": {
    "granola": {
      "type": "stdio",
      "command": "node",
      "args": ["/path/to/granola-mcp/dist/mcp/server.js"],
      "env": {
        "GRANOLA_DATA_DIR": "/path/to/granola-mcp/export",
        "OPENAI_API_KEY": "sk-..."
      }
    }
  }
}

Additional configuration and usage notes

The server's environment variables include the data directory for Granola export data and your OpenAI API key. If you manage multiple workspaces or data directories, you can customize the GRANOLA_DATA_DIR for each MCP configuration to point to the appropriate export folder.

Tools and endpoints you can use with this MCP server

The MCP server exposes a set of tools you can call from your client to interact with your Granola data. Each tool returns structured results such as summaries, quotes, and document lists.

Security considerations

Protect your OpenAI API key and restrict access to the MCP server. Do not expose the server to untrusted networks. If you automate starting the server, secure your environment variables and rotate keys as needed.

Troubleshooting tips

If the server fails to start, verify that the paths in the configuration exist, the Granola export directory contains data, and the API key is valid. Check that Node.js 18+ is installed and that you have network access to any required services.

Available tools

search

Semantic search across meetings, returns summaries plus quotes with speaker attribution.

search_themes

Find documents by theme such as pain-points, feature-requests, and decisions.

list_folders

List all folders with document counts to understand data structure.

list_documents

List documents with brief summaries for quick review.

get_document

Get full document details including all themes and quotes.

get_transcript

Retrieve the raw transcript data when needed.

get_themes

List available themes with definitions.