Workflowy Desktop MCP

Connect an external AI tool using the MCP server built into the Workflowy desktop app.

View as Markdown

Workflowy Desktop MCP lets AI tools running on your computer read, search, navigate, and update your Workflowy.

When to use

Use Desktop MCP when you want an AI tool to work directly with your Workflowy content. It's useful for searching across your notes, summarizing project notes, adding tasks from an AI chat, and reorganizing nodes.

No API key is required, and the desktop app authenticates using your signed-in Workflowy account.

Security note: Workflowy generates a unique Connection Link for your AI tool. Treat it like a private key and only paste it into AI tools you trust. You can disable Local MCP or close the Workflowy desktop app at any time to stop access.

Steps to connect

Before you start, make sure you have an AI tool that supports local MCP servers over Streamable HTTP, such as Codex, Claude Code, Cursor, or VS Code.

  1. Download or update the Workflowy desktop app
  2. Open the Workflowy desktop app and sign in.
  3. Open the More Menu and select Settings.
  4. Scroll down to the MCP section.
  5. Enable Local MCP.
  6. Copy the generated Connection Link.
  7. Add that link to your AI tool's MCP settings using Streamable HTTP.
  8. Your AI tool can now access your Workflowy notes.
  9. Keep the Workflowy desktop app running while you use the connection.

Examples

Your AI tool can use MCP to read, search, and update your Workflowy notes, which opens up many useful ways to work with AI. Here are a few examples:

  • Search your notes on a specific topic
  • Ask your AI to summarize notes under a project node
  • Store the output of an AI chat in a note for safekeeping and reference
  • Add tasks to your today list from an AI chat
  • Move completed tasks into an Archive node
  • Attach files to a project node in Workflowy and have your AI read and analyze them
  • Use Workflowy to control your AI assistant. For example, ask your AI to watch for a specific #tag. When you add that #tag to a task in Workflowy, your AI can work on it and report back.

Troubleshooting

  • If tools do not appear, restart your AI tool after adding the Connection Link.
  • If the AI tool cannot connect, make sure the Workflowy desktop app is open and Local MCP is enabled.
  • If Workflowy says MCP could not start, another app may be using the default port. Change the port in the MCP settings and copy the new Connection Link.
  • If you switch Workflowy accounts in the desktop app, the MCP server follows the currently signed-in account.
Up next
Run your own Workflowy connector
Compare local MCP and command-line options for connecting external AI tools to Workflowy.