mcp -cloudflare
MCP server from cloudflare/mcp-server-cloudflare
claude mcp add --transport stdio cloudflare-mcp-server-cloudflare npx mcp-remote https://bindings.mcp.cloudflare.com/mcp \ --env MCP_REMOTE_VERBOSE="optional: set to 1 for verbose output"
How to use
This repository provides Cloudflare-based MCP servers that you can access from any MCP client to manage Cloudflare services via natural language. The servers expose a remote MCP interface (via the streamable-http transport at /mcp and a deprecated SSE transport at /sse) so clients can read configurations, inspect data, receive suggestions, and even execute changes across Cloudflare products. You can connect to these servers either directly using the remote configuration mechanism (for example with mcp-remote) or by running the provided example configurations in your client. The two example endpoints shown are cloudflare-observability and cloudflare-bindings, which allow you to query observability data and manage Workers Bindings respectively through natural language prompts. When using these remote servers, your MCP client will send natural language prompts, and the server will translate those into Cloudflare API calls, returning results or performing actions as requested.
How to install
Prerequisites:
- Node.js (recommended: LTS version, e.g., 18.x or newer) and npm installed on your machine.
- Optional: npx is included with npm, so you can run remote MCP servers without a global install.
Installation steps:
-
Ensure Node.js and npm are installed.
- Check: node -v and npm -v
-
Install the mcp-remote tool globally (optional but convenient) or use npx directly.
- Global install: npm install -g mcp-remote
- Or you can run without global install using npx as shown in the examples below.
-
Create a local MCP configuration file using the endpoints provided in this repo’s README. Example (save as mcp-config.json):
{ "mcpServers": { "cloudflare-observability": { "command": "npx", "args": ["mcp-remote", "https://observability.mcp.cloudflare.com/mcp"] }, "cloudflare-bindings": { "command": "npx", "args": ["mcp-remote", "https://bindings.mcp.cloudflare.com/mcp"] } } }
-
Run the remote servers from your client or environment:
- With npx (no global install): npx mcp-remote https://observability.mcp.cloudflare.com/mcp npx mcp-remote https://bindings.mcp.cloudflare.com/mcp
- If you installed mcp-remote globally: mcp-remote https://observability.mcp.cloudflare.com/mcp mcp-remote https://bindings.mcp.cloudflare.com/mcp
-
In your MCP client, point to the server URLs above or load the mcp-config.json as your remote servers configuration.
Note: This repository primarily provides remote MCP endpoints. If you want to host these servers yourself, ensure you have access to Cloudflare APIs and proper credentials in your environment as described in the repository’s contributing and troubleshooting sections.
Additional notes
Tips and known considerations:
- Transport options: The servers support the streamable-http transport via /mcp and the SSE transport (deprecated) via /sse. Prefer /mcp for new clients.
- API tokens: Some MCP servers require Cloudflare API tokens with appropriate permissions for the specific services (e.g., observability, bindings). Ensure tokens are scoped correctly in your client or environment.
- If Claude or other LLMs encounter context-length issues, consider breaking up requests into smaller steps or limiting the number of tools invoked per prompt to reduce context usage.
- When using OpenAI's Responses API, ensure your API token has the required permissions for the Cloudflare MCP server you target.
- If you plan to extend functionality, this repo hosts multiple servers (e.g., observability, bindings, etc.). Each server may have its own set of capabilities and required permissions.
- SSE transport is deprecated; prefer the /mcp endpoint with streamable-http for better reliability and performance.
Related MCP Servers
iterm
A Model Context Protocol server that executes commands in the current iTerm session - useful for REPL and CLI assistance
mcp
Octopus Deploy Official MCP Server
furi
CLI & API for MCP management
editor
MCP Server for Phaser Editor
DoorDash
MCP server from JordanDalton/DoorDash-MCP-Server
mcp
MCP сервер для автоматического создания и развертывания приложений в Timeweb Cloud