Notion
About
Connects AI assistants to your Notion workspace, allowing you to search, create, and manage content using natural language.
Details
- Author
- shameerpc5029
- Categories
- Productivity, Other, Project Management, Knowledge Base
Jump to
Setup
Install Notion in your MCP client (Claude Desktop, Cursor, Windsurf, and others).
Repository: https://github.com/shameerpc5029/notion-mcp
Follow the installation instructions in the repository README, then restart your MCP client.
A powerful Model Context Protocol (MCP) server that connects AI assistants like Claude to your Notion workspace. Seamlessly search, create, and manage your Notion content through natural language conversations!
Transform your AI assistant into a Notion powerhouse! With this MCP server, you can:
- πSearchthrough all your Notion pages and databases
- πCreatenew pages, databases, and content
- βοΈUpdateexisting pages and properties
- πQuerydatabases with complex filters
- πManageyour entire Notion workspace through conversation
- Python 3.8 or higher
- A Notion workspace
- Nango account (for OAuth authentication)
- Claude Desktop app (for AI integration)
# Clone or download the files git clone <your-repo-url> cd notion-mcp-server # Install dependencies uv sync
- β
Read content
- β
Insert content
- β
Update content
- Open any Notion page you want to access
- Click "Share" β "Invite"
- Select your integration
- Repeat for all pages/databases you want to use
- Create Nango Account:https://nango.dev
- Add Notion Integrationin your Nango dashboard
- Configure OAuthwith your Notion integration credentials
- Get your Nango credentials(Base URL, Secret Key, Connection ID)
Create a.envfile with your credentials:
# Nango OAuth Configuration NANGO_BASE_URL=https://api.nango.dev NANGO_SECRET_KEY=your_nango_secret_key_here NANGO_CONNECTION_ID=your_connection_id_here NANGO_INTEGRATION_ID=notion # Optional: Direct Notion Token (if not using Nango) # NOTION_TOKEN=your_direct_notion_token_here
Add this to your Claude Desktop configuration file:
On macOS:~/Library/Application Support/Claude/claude_desktop_config.jsonOn Windows:%APPDATA%\Claude\claude_desktop_config.json
{ "mcpServers": { "notion": { "command": "uvx", "args": ["git+https://github.com/Shameerpc5029/notion-mcp.git"], "env": { "NANGO_BASE_URL": "https://api.nango.dev", "NANGO_SECRET_KEY": "your_nango_secret_key", "NANGO_CONNECTION_ID": "your_connection_id", "NANGO_INTEGRATION_ID": "notion" } } } }
# Test the server directly python notion-mcp-server.py # Or restart Claude Desktop and try asking: # "Search my Notion for meeting notes" # "Create a new task in my project database"
Once configured, you can have natural conversations like:
You: "Show me all my incomplete tasks" Claude: [Queries your tasks database and shows open items] You: "Create a new task called 'Review MCP integration' due next Friday" Claude: [Creates the task with proper due date] You: "Mark the first task as completed" Claude: [Updates the task status to Done]
You: "Create a meeting notes page for today's standup" Claude: [Creates a new page with meeting template] You: "Add action items from our discussion to that page" Claude: [Appends bullet points with the action items]
You: "Search for all pages mentioning 'API documentation'" Claude: [Finds and lists relevant pages] You: "Create a new database to track our API endpoints" Claude: [Creates a structured database with relevant properties]
- Check your.envfile has correct Nango credentials
- Verify your Nango integration is active
- Ensure your Notion integration has proper capabilities
- Make sure you've shared the page/database with your integration
- Check the page/database ID is correct
- Verify the integration has read access
- Check the absolute path inclaude_desktop_config.jsonis correct
- Ensure Python is in your system PATH
- Try running the server manually first to test
- Verify Nango connection is working
- Check yourNANGO_CONNECTION_IDmatches your actual connection
- Try refreshing your Nango integration
from notion_mcp_server import NotionClient client = NotionClient() user = client.get_current_user() print(f"Connected as: {user.get('name')}")
- Never commit your.envfileto version control
- Keep your Nango secret key secure
- Only share Notion pages that the integration needs access to
- Regularly review your integration's access in Notion settings
notion-mcp/ βββ main.py # Main MCP server βββ pyproject.toml # Python dependencies βββ .env.example βββ .env # Your credentials (don't commit!) βββ README.md # This file
- Check the logs- Look for error messages in Claude Desktop or terminal
- Verify permissions- Ensure your integration can access the pages
- Test manually- Run the server directly to isolate issues
- Check Notion status- Sometimes Notion API has outages
- Try creating complex database queries
- Set up automated content creation workflows
- Use Claude to help organize and restructure your Notion workspace
- Explore advanced filtering and sorting options
This project is open source. Feel free to modify and adapt for your needs!
Made with β€οΈ for the AI-powered productivity enthusiasts
An MCP server for interacting with your Notion workspace, enabling LLMs to manage pages and databases.
Official Notion MCP server for searching, reading, creating, and updating Notion pages, databases, and workspace content from AI agents.
Markdown-first Notion MCP server with 9 composite tools, 39 actions, and ~77% token reduction via tiered docs.
Connect your AI assistant to your personal knowledge base. Search, save links, create notes and to-dos. AI processes everything automatically.
Connects AI assistants to your Notion workspace to search and manage pages, databases, and content.
Interact with Notion's API to read, create, and modify content using natural language.
Manage and interact with your entire Notion workspace.
Search, read, create, and update pages and databases in your Notion workspace.
Integrate with Notion workspaces to manage databases, pages, and content.
Interact with Notion's API to manage todo lists, databases, and content organization.
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.





