Any local MCP server, any agent harness.

webmcp.fast is a hosted gateway for the Model Context Protocol. You run a small open-source daemon on your own machine, attach the local MCP servers you choose, and each one gets a stable, authenticated URL that a cloud agent can use as a remote MCP server:

https://<handle>.webmcp.fast/<device>/<server>/mcp

Point Claude, ChatGPT, Grok, Cursor or your own agent harness at that URL and it sees exactly that server's tools, resources and prompts, as if the server were hosted. Your API keys, files, repositories and command-line tools stay on your machine. Only MCP messages cross the wire.

How it works

• The daemon dials out. It holds one outbound WebSocket to the gateway. Your machine opens no inbound port and needs no tunnel software, VPN or Cloudflare account of its own.
• One URL per server. There is no aggregate endpoint. A connector reaches one server on one device and nothing else.
• Standard OAuth 2.1. Cloud agents connect the way they connect to any remote MCP server: dynamic client registration, PKCE, and a consent screen where you approve that one endpoint. A revocable connector token is the fallback for harnesses without OAuth.
• You decide what is exposed. Servers are added only from your machine with webmcp attach. The dashboard can disable or remove a server. It can never add one or run a command.

Set it up by asking your coding agent

Say to a local coding agent: "set up webmcp.fast for my filesystem server". The agent does the rest, apart from two things only you can do.

1. Install the daemon. npm i -g @proticom/webmcp, or skip the install and run npx @proticom/webmcp up in the next step. npm works on macOS, Linux and Windows and needs no compiler: the package carries prebuilt binaries. On Linux the GitHub release tarballs also work: curl -fsSL https://github.com/proticom/webmcp/releases/latest/download/install.sh | sh. The macOS tarballs are unsigned, so use npm on a Mac. The daemon is open source.
2. Run webmcp up. It generates a device key on the machine and prints one link, https://webmcp.fast/activate?code=XXXX-XXXX. Show that link to the human and wait; webmcp up returns when they have approved. It also offers any MCP servers it finds in existing Claude, Cursor, Codex and VS Code configs.
3. The human opens the link in their own browser. They sign in with the 6-digit code emailed to them (this creates the account if they have none), pick a handle, and approve the device. The code is never given to the agent: do not ask for it, and do not offer to type it.
4. Attach the server. webmcp attach <alias> -- <command> [args...] for a stdio server, or webmcp attach <alias> --http <local url> for one already listening on localhost. The alias is 1 to 32 characters of a-z, 0-9 and -.
5. Keep it online. webmcp service install keeps the daemon running in the background: a launchd agent on macOS, a systemd user unit on Linux (run loginctl enable-linger $USER so it survives logout). Windows has no service; run webmcp connect under your own supervisor there.
6. Hand over the connector URL. Print https://<handle>.webmcp.fast/<device>/<server>/mcp with the real handle, device and alias. The human pastes it into Claude, ChatGPT or Grok as a custom connector, signs in, and clicks Allow.

Two steps are deliberately left to the human, and no agent can do them:

• Proving they own the email address. The 6-digit code goes to their inbox and is typed into their own browser. It is bound to that browser, so a code read aloud to an agent is useless anywhere else.
• Approving each cloud agent. Every connector goes through an OAuth consent screen that names the one endpoint being granted. The human clicks Allow, once per agent per server.

The full walkthrough for agents is at https://webmcp.fast/docs/agents. A cloud chat agent with no shell can start the same setup through the management MCP server at https://webmcp.fast/mcp.

Free tier

Free, self-serve, no credit card: 1 handle, 1 device, 1 server, 1 connector. Sign up by signing in with your email address. A paid tier for more devices, servers and connectors is planned and not yet available. See pricing.

Learn more

• Quickstart: from nothing to a working connector.
• Agent-driven setup: what an agent should do, step by step.
• Security and the security model: how exposure is limited.
• Pricing, about, contact, privacy, terms.
• For agents: llms.txt, OpenAPI, MCP server card, agent skill.