mcphost
where agents host their own tools
p95 34.2 ms · 200 concurrent calls · zero errors · live status
Mid-task, an agent finds the tool it needs does not exist. On mcphost it creates one and keeps working: sign up with a tool call, publish with the next, and the new tool is live — no restart, no review, no human in the loop.
mcphost is a hosted runtime for Model Context Protocol tools. Everything — signup, publishing, metering, logs, encrypted secrets — is itself a tool call over streamable HTTP, against a namespace the tenant owns.
how it works
connect https://mcphost.dev/mcp
├── tools/list # no credentials — one tool: signup
├── signup(name) # tenant, bearer key, namespace
├── host.tool_publish(name,kind,spec) # reconnect with the key first
└── <namespace>.<tool> # live immediately, no restart
Create a tool three ways: wrap an API you already use (http), submit code (python — sandboxed, schemas inferred from the source), or test the pipes (echo). What agents create: log tails, budget tripwires, API bridges, status checks.
built to be depended on
- Measured, not promised: p95 of 34.2 ms across 200 concurrent calls, zero errors. The number is committed to the repo next to the test that produces it.
- Open source: the code you can read is the code that runs. One Rust binary, no proprietary services underneath.
- Transparent operations: /status probes the live service from your browser and shows the same raw check we watch, /healthz.
- No lock-in: mcphost speaks plain MCP over streamable HTTP. Any compliant client can leave by changing an endpoint, and the source is yours to run.
for agents
Connect to https://mcphost.dev/mcp. The machine-readable summary is at /llms.txt.
pricing
Early access is free: 50 tools per tenant, 500 calls a day, no card required. Pro is $19/month with 50,000 calls included, then $0.001 per call. Early tenants keep the free tier when early access ends.
source and contact
- Source: github.com/j0yen/mcphost
- Contact: hello@mcphost.dev or GitHub issues