claudekit / tools / notion
[ MCP · Data Engineering ]

Notion MCP

Notion's official remote MCP server. Read and write Notion workspaces directly from Claude Code — no self-hosting required.

published ·updated
$ claude mcp add --transport http notion https://mcp.notion.com/mcp copy

Official Notion remote MCP server — hosted by Notion

What it does

Notion hosts an official remote MCP server at https://mcp.notion.com/mcp. Connecting Claude Code to it gives the agent read/write access to your Notion workspace — pages, databases, blocks, comments — authenticated through Notion’s OAuth flow.

How to connect

1. Claude Code CLI

claude mcp add --transport http notion https://mcp.notion.com/mcp

After adding, run /mcp inside Claude Code to get the OAuth link. Sign in with Notion in the browser to finish the connection.

2. Connect from Notion settings

You can also connect directly from Notion → Settings → Integrations → Notion MCP.

Scope options

OptionDescription
--scope local (default)Available only in the current project
--scope projectShared across the team via .mcp.json
--scope userAvailable across all your projects

Features

  • Search pages — full-workspace content search
  • Write documents — auto-generate PRDs, technical specs, release notes
  • Database queries — inspect and update project status
  • Edit content — markdown-based page updates

Use cases

  1. Search planning docs to confirm requirements, then update the Notion task state automatically after implementation.
  2. Generate release notes in Notion from code changes.
  3. Query a Notion project database for an at-a-glance status of ongoing work.

Why remote MCP

No local server to run, no token to rotate yourself — Notion handles hosting and auth. Ideal when you want the integration to “just work” without ops overhead.

Notes

  • Image and file upload aren’t supported yet.
  • OAuth is required, so fully headless/automated environments aren’t supported.
  • Use /context to check current token usage.
  • Also available via the Claude Code plugin marketplace, which ships with bundled skills and slash commands.
§ 7

See also

same category · curated
[01]
[Tool] Firecrawl · 🔥 The API to search, scrape, and interact with the web for AI. Three integrated capabilities — Search, Scrape, Interact — exposed through one API. Open source under AGPL-3.0 and self-hostable via docker-compose, the engine also powers the firecrawl.dev cloud SaaS run by the same team.
tool · claudekit.io / tools / firecrawl
[02]
[MCP] Firecrawl MCP · 🔥 Official Firecrawl MCP Server — Adds powerful web scraping and search to Cursor, Claude and any other LLM clients. Exposes 12+ tools spanning single-page scrape, batch processing, site crawl, search, structured extraction, autonomous research agent, and interactive page automation, returning clean LLM-ready markdown.
tool · claudekit.io / tools / firecrawl-mcp
[03]
[MCP] Perplexity MCP · The official MCP server implementation for the Perplexity API Platform. Provides AI assistants with real-time web search, reasoning, and research capabilities through Sonar models and the Search API.
tool · claudekit.io / tools / perplexity-mcp
§ 8

Frequently Asked Questions

frequently asked
§ 8.1
What is Notion MCP?
Notion's official remote MCP server. From Claude Code you can search, read, and write pages and databases in your Notion workspace.
§ 8.2
Where does it run?
Claude Code and Claude Desktop.
§ 8.3
How do I connect it?
Run `claude mcp add --transport http notion https://mcp.notion.com/mcp`, then `/mcp` to get the OAuth link and sign in with Notion. You can also connect from **Notion → Settings → Integrations → Notion MCP**.
§ 8.4
How do I share it with my team?
Use `--scope project` to share via `.mcp.json` across the team. `--scope user` enables it across all your projects; the default `--scope local` keeps it project-local.
§ 8.5
What can it do?
Full-workspace page search, auto-generation of PRDs, technical specs, and release notes, database queries and updates, and markdown-based page editing.
§ 8.6
Are there limitations?
Image and file uploads aren't supported yet, and OAuth authentication means fully headless/automated environments aren't supported.