We built a cloud platform that lets you deploy MCP servers (agents, tools, ChatGPT apps) to production instantly. Think Vercel, but for agents using the Model Context Protocol (MCP).
The problem: MCP servers are great locally, but running them in production is a pain. You need hosting, handle auth, manage secrets, deal with long-running processes, and connect everything securely.
With mcp-c, it's: uvx mcp-agent deploy
You get a production URL that works with Claude Desktop, ChatGPT, Cursor, or any MCP client.
What makes this different: - Durable execution (agents can run for hours without timing via Temporal) - Built-in secrets management - Full MCP spec support (sampling, notifications, logging) - Free during beta
Try this hosted ChatGPT Pizza example:
Pizza ordering agent: https://pizzaz.demos.mcp-agent.com/sse
We're the team behind mcp-agent (8K GitHub stars), and this is our answer to "how do I actually ship this?"
Open beta, free to use. Would love your feedback on what would make this more useful.
Link to docs: https://docs.mcp-agent.com/get-started/cloud