rclone
MCP server for rclone.
claude mcp add --transport stdio rclone-ui-rclone-mcp npx -y rclone-mcp \ --env RCLONE_URL="http://localhost:5572"
How to use
This MCP server exposes the Rclone RC API through an AI-friendly interface. It auto-generates a suite of tools from the rclone-openapi spec via the rclone-sdk, providing a large set of endpoints (about 98 tools) grouped into toolsets such as core, config, operations, and sync. By running the server with the rclone-mcp package, you enable natural-language capable assistants to list directories, copy or move files, manage remotes, and perform cloud-storage operations through the RC API. The server can be accessed locally via the RC daemon URL you provide (default http://localhost:5572) and exposed to clients through standard MCP transports (stdio or HTTP).
How to install
Prerequisites:
- Node.js and npm (or an environment that can run npx)
- A running rclone remote control daemon (rcd) accessible via HTTP, configured with no-auth or with HTTP Basic Auth as desired
Install and run (stdio transport):
-
Ensure rclone rcd is running, e.g.: rclone rcd --rc-no-auth
or with auth:
rclone rcd --rc-user=admin --rc-pass=secret
-
Run the MCP server using npx (no global install required): npx -y rclone-mcp
-
If you need HTTP transport instead, you can run the HTTP variant: npx rclone-mcp http --port 3000
Docker-based run (build-and-run):
- Build the image (Dockerfile should be provided by the project): docker build -t rclone-mcp .
- Run the container, passing the RCLONE_URL to the RC daemon: docker run -i --rm -e RCLONE_URL=http://host.docker.internal:5572 rclone-mcp
Environment and configuration:
- Set RCLONE_URL to point to your rclone rc daemon (default http://localhost:5572).
- Optional: RCLONE_USER and RCLONE_PASS if you’re protecting the RC API with Basic Auth.
- You can control exposed toolsets with RCLONE_TOOLSETS or via command-line flags such as --toolsets and --read-only.
Additional notes
Tips and caveats:
- The server exposes a large set of tools; use the default toolsets (core, config, operations, sync) for a broad capability set. Use --toolsets to restrict exposed tools if you want a focused interface.
- In read-only mode (RCLONE_READ_ONLY=1 or --read-only), mutating operations like copy/move/delete and certain config actions are hidden. This is useful for AI safety in untrusted environments.
- If you’re exposing the server over HTTP, consider securing with authentication, rate limiting, and TLS as appropriate for your deployment.
- When using Docker, ensure the container can access the host's rclone rc daemon; you may need to map the RC URL accordingly (e.g., http://host.docker.internal:5572).
- The RCLONE_TOOLSETS environment variable supports values like default, all, or a comma-separated list of toolsets to tailor the available commands.
Related MCP Servers
awesome-claude-skills
A curated list of awesome Claude Skills, resources, and tools for customizing Claude AI workflows
awesome-agent-skills
A curated list of skills, tools, tutorials, and capabilities for AI coding agents (Claude, Codex, Antigravity, Copilot, VS Code)
dbhub
Zero-dependency, token-efficient database MCP server for Postgres, MySQL, SQL Server, MariaDB, SQLite.
Unity
AI-powered bridge connecting LLMs and advanced AI agents to the Unity Editor via the Model Context Protocol (MCP). Chat with AI to generate code, debug errors, and automate game development tasks directly within your project.
coplay-unity-plugin
Unity plugin for Coplay
mcpx
Token-efficient MCP client: TypeScript schemas instead of JSON, LLM-friendly syntax, batch calls, TOON output. Built for Claude/GPT automations.