Granola MCP Server
Integrates Granola.ai meeting intelligence with MCP clients to provide direct access to meeting notes, transcripts, and workspace organization. It allows users to search and retrieve meeting content through natural language by leveraging existing Granola credentials.
README
Granola MCP Server
A Model Context Protocol (MCP) server for integrating Granola.ai meeting intelligence with Claude Desktop and other MCP clients. Access your meetings, notes, and transcripts directly from AI assistants.
⚠️ Disclaimer
This is an unofficial, community-built tool. It is not affiliated with, endorsed by, or supported by Granola, Inc.
This server uses undocumented APIs discovered through network inspection. Use of this tool may violate Granola's Terms of Service. Use at your own risk.
If you are from Granola and would like this repository removed or would like to discuss official API access, please open an issue.
Features
- 🔍 Meeting Search - Search meetings by title, content, and participants
- 📝 Full Content Access - Retrieve complete meeting notes in Markdown format
- 🏢 Workspace Management - List and filter by workspaces (organizations)
- 📁 Folder Organization - List and filter by document folders
- 🔄 Real-time API Access - Direct access to Granola's API
- 🔐 Secure Authentication - Uses your existing Granola credentials
- ⚡ Smart Caching - 5-minute cache for optimal performance
Prerequisites
- Node.js 18+
- Granola.ai installed and logged in on macOS
- Claude Desktop (or another MCP client)
Installation
git clone https://github.com/EoinFalconer/granola-mcp-server.git
cd granola-mcp-server
npm install
npm run build
Configuration
Claude Desktop
Add to your claude_desktop_config.json:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
{
"mcpServers": {
"granola": {
"command": "node",
"args": ["/path/to/granola-mcp-server/index.js"]
}
}
}
Then restart Claude Desktop.
Available Tools
| Tool | Description | Parameters |
|---|---|---|
search_meetings |
Search meetings by title/content | query, limit? |
get_meeting_details |
Get meeting metadata | meeting_id |
get_meeting_content |
Get full notes as Markdown | meeting_id |
list_workspaces |
List all workspaces | - |
list_folders |
List document folders | workspace_id? |
filter_by_workspace |
Filter meetings by workspace | workspace_id |
filter_by_folder |
Filter meetings by folder | folder_id |
Usage Examples
Once configured, use natural language in Claude:
- "Search for meetings about quarterly planning"
- "Get the content from yesterday's standup"
- "List all my workspaces"
- "Show me meetings in the Sales folder"
How It Works
Authentication
The server reads your existing Granola credentials from:
~/Library/Application Support/Granola/supabase.json
No additional login required - it uses your existing Granola session.
API Integration
- Direct REST API calls to Granola's backend
- ProseMirror content converted to readable Markdown
- In-memory caching with 5-minute TTL
Troubleshooting
"Failed to get access token"
Ensure Granola is installed and you're logged in. The server needs:
~/Library/Application Support/Granola/supabase.json
Server not appearing in Claude
- Verify the config path is correct
- Run
npm run build - Restart Claude Desktop completely
Development
npm run build # Compile TypeScript
npm run watch # Watch mode
npm run dev # Build and run
Credits & Acknowledgments
This project was inspired by and built upon the reverse-engineering work of:
- Joseph Thacker - Detailed blog post on reverse-engineering Granola's API
- getprobo/reverse-engineering-granola-api - API documentation and research
Thank you for sharing your findings with the community! 🙏
Related Projects
- cobblehillmachine/granola-claude-mcp - Python-based, cache file approach
- btn0s/granola-mcp - TypeScript, similar API approach
Contributing
Contributions welcome! Please feel free to submit issues and pull requests.
License
MIT © Eoin Falconer
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。