Home / MCP / Notion ReadOnly MCP Server
Provides read-only access to Notion content for fast AI-assisted queries via a minimal MCP surface.
Configuration
View docs{
"mcpServers": {
"notion_readonly_npx": {
"command": "npx",
"args": [
"-y",
"notion-readonly-mcp-server"
],
"env": {
"OPENAPI_MCP_HEADERS": "{\"Authorization\": \"Bearer ntn_****\", \"Notion-Version\": \"2022-06-28\" }"
}
}
}
}You can run a focused, read-only MCP server that exposes essential Notion data retrieval endpoints with high performance. This server is optimized for AI assistants to query Notion content quickly, using a minimal, safe surface area designed for fast responses and easy integration.
You connect your MCP client to the Notion ReadOnly MCP Server to access pages, blocks, databases, comments, and page properties. Use the provided MCP endpoints to retrieve specific content or to fetch a complete page with all related content in a single call. The design emphasizes speed and a small tool surface to keep AI assistant reasoning fast and reliable.
Prerequisites: ensure you have Node.js and npm installed on your machine, and Docker if you plan to run the Docker configuration. You also need access to a Notion integration with read permissions.
{
"mcpServers": {
"notion_readonly_npx": {
"command": "npx",
"args": ["-y", "notion-readonly-mcp-server"],
"env": {
"OPENAPI_MCP_HEADERS": "{\"Authorization\": \"Bearer ntn_****\", \"Notion-Version\": \"2022-06-28\" }"
}
},
"notion_readonly_docker": {
"command": "docker",
"args": [
"run",
"--rm",
"-i",
"-e", "OPENAPI_MCP_HEADERS",
"taewoong1378/notion-readonly-mcp-server"
],
"env": {
"OPENAPI_MCP_HEADERS": "{\"Authorization\":\"Bearer ntn_****\",\"Notion-Version\":\"2022-06-28\"}"
}
}
}
}Security and access You can limit workspace access by configuring the integration’s capabilities to read content only. Use a read-only integration token for safer operation.
Connect your content to the integration Make sure relevant pages and databases are connected to your integration. Go to the Notion page, open the options, and connect to the integration so the MCP server can access the content.
Usage patterns The server exposes 7 essential tools for retrieving content. These are designed to be used in AI-assisted workflows where quick access to pages, blocks, databases, and comments is required, including the ability to fetch a whole page with its related content in one call.
Recipes and examples emphasize practical retrieval: request page details, fetch block children with parallel processing, retrieve databases and page properties, or obtain a full page with the one-pager tool. The aim is to deliver fast, reliable data to AI assistants with minimal tool surface.
Get page information including ID, title, type, and metadata for a Notion page.
Fetch the content blocks under a page or block, with support for parallel processing to speed up retrieval.
Retrieve details about a specific block, such as its type, content, and formatting.
Get information about a Notion database, including its properties and schema.
Get comments associated with a page or block.
Fetch a specific property from a Notion page.
Recursively retrieve a full Notion page with all blocks, databases, comments, and properties in one call