--- { "title": "Run your own Workflowy connector", "intro": "Compare local MCP and command-line options for connecting external AI tools to Workflowy." } --- Run your own connector when you want capabilities beyond the connection built into the Workflowy desktop app, such as a persistent local cache, multiple accounts, command-line workflows, or deeper automation. # Choose a connector
Connector Best for Setup Platforms
Workflowy Local MCP app A visual desktop app, fast local search, backups, and several MCP clients Install the app and add a Workflowy API key macOS, Windows, Linux
Workflowy for Claude Desktop The shortest setup for Claude Desktop and multiple Workflowy accounts Install a .mcpb extension and add one or more API keys macOS, Windows
Workflowy CLI Terminal use, scripts, automations, and a configurable MCP server Install wf, sign in, and run wf mcp macOS, Windows, Linux
All three options run locally and use a Workflowy API key. Each project is open source, so advanced users can inspect the source and adapt the setup. # When built-in Desktop MCP is enough If you only need to connect one compatible AI tool on the same computer, start with [Workflowy Desktop MCP](/help/mcp-server). It uses your signed-in desktop session and does not require a separate connector or Workflowy API key. # Community connectors Other community-built Workflowy connectors exist. Before installing one, check how it authenticates, which operations it can perform, whether it stores a local copy of your data, and how recently it has been maintained. Community connectors are not maintained or supported by Workflowy unless a guide says otherwise.