How does this differ from the official Notion MCP server?
The official server returns large JSON blobs; this server returns Markdown, which is more efficient for LLMs to process and understand.
@larryhudson
基本情報
Notion MCP server that uses Markdown for better LLM support
以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。
{
"mcpServers": {
"simple-notion-mcp-server": {
"command": "npx",
"args": [
"@modelcontextprotocol/inspector",
"npx",
"tsx",
"src/index.ts"
]
}
}
}ツール一覧はまだありません
この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。
上のボタンで一時的なサンドボックスでこのサーバーを実行し、ツール一覧を取得できます。
Simple Notion MCP Server is a Model Context Protocol (MCP) server that bridges AI assistants with the Notion API. Unlike the official Notion MCP server, which returns large JSON blobs, this server renders Notion pages as clean Markdown, making them more efficient for LLMs to process and understand.
Create a Notion integration at https://www.notion.so/my-integrations and obtain an API token. Set the token as the NOTION_TOKEN environment variable or in a .env file. Share any Notion pages you want to access with your integration. Then configure your MCP-compatible application (VS Code, Claude, etc.) to run the server via npx -y @larryhudson/simple-notion-mcp-server with the token. Use the get_page tool with a page_id parameter to retrieve and render a page as Markdown.
The official server returns large JSON blobs; this server returns Markdown, which is more efficient for LLMs to process and understand.
It runs via `npx` (Node.js) and requires a Notion API token. No additional installation is needed beyond Node.js.
Data is fetched from the Notion API using your integration token. Pages must be shared with the integration. No data is stored locally.
Currently only supports retrieving pages (not updating or creating). Image handling is basic and may be improved in future updates (e.g. rendering as base64 image blocks).
Set the `NOTION_TOKEN` environment variable to your Notion integration API token. The server uses this token to authenticate with the Notion API.
Project context your AI can query through MCP — with source evidence.

One permissioned folder your team and your AI agents both log into.
Memory for AI agents that stays correct as facts change. Open source and local first: one memory for Claude Code, Codex, Cursor and your own agents.

The Cloud is the AI-native workspace that gives every AI — Claude, ChatGPT, Gemini, and other Frontier & Free or Local models — shared memory and real agents. Not cloud computing: thecloud.so is one workspace where your

Exact, shared memory for AI agents: a hierarchical database where totals always add up, every write is attributed to the agent and person behind it, and any figure traces back to its source file. Agents create databases,
コメント