mcp.hosting — Deploy MCP Servers Without the Infrastructure Headache

MCP servers are easy to build. Hosting them is not.

The Model Context Protocol has made it straightforward to give AI assistants access to your tools, data, and APIs. Writing an MCP server takes an afternoon. But then you need to actually run it somewhere.

That means provisioning infrastructure, setting up health checks, handling TLS, managing authentication, monitoring uptime, and dealing with the operational overhead that comes with any production service. For a protocol designed to make AI tooling simple, the deployment story has been anything but.

mcp.hosting fixes this. It is a managed hosting platform purpose-built for MCP servers.

What mcp.hosting gives you

• One-click deployment. Push your MCP server and it is live. No Dockerfiles, no Kubernetes manifests, no load balancer configuration.
• Auto-provisioned API keys. Users connecting to your hosted MCP server get credentials automatically. No manual key management.
• Session proxy. The session proxy spec handles stateful MCP connections reliably, maintaining session state across reconnections without your server needing to manage it.
• Uptime monitoring. Built-in health checks and status tracking. Know when your server is degraded before your users do.
• Team management. Role-based access, audit logs, and multi-team support. Control who can deploy, configure, and access your MCP servers.
• GitHub integration. Connect your repo and get automatic deployments on push. Your MCP server stays in sync with your codebase.
• Custom domains. Bring your own domain with automatic DNS and TLS configuration.

Validate before you deploy

Before deploying to mcp.hosting, you can run your server through mcp-compliance to verify it meets protocol standards. This catches configuration errors, missing capabilities, and spec violations before they hit production.

Combined with ctxlint for validating your MCP client configs, you get a complete validation pipeline from development to deployment.

Track costs with Token Meter

Token Meter is available on mcp.hosting as a hosted MCP server. Add it to your client config and your AI assistant can check LLM spend, compare model prices, and manage budgets — all without self-hosting anything.

If you are building MCP servers that make LLM calls, Token Meter gives you visibility into what those calls cost. The combination of mcp.hosting for deployment and Token Meter for cost tracking covers both sides of the operational equation.

Need private infrastructure instead?

Not every MCP server should be on the public internet. If your servers need to stay on a private network, tailscale-mcp lets you manage your tailnet directly from your AI assistant and provides secure access to MCP servers without exposing endpoints publicly. Use mcp.hosting for your public-facing servers and tailscale-mcp for everything that needs to stay private.