Engress

Engress gives your local AI models a public HTTPS URL. You run a small agent on your machine; it dials outbound to the Engress edge pool. There is no port forwarding, no exposed listener on your router, and no inbound firewall rule to maintain.

Tools like Cursor, notebooks, and teammates' scripts call your model through that URL. Traffic streams back through the tunnel to whatever you pointed the agent at — Ollama on port 11434, LM Studio on 1234, or any local HTTP server.

How it works in three steps

1. Install the agent on the machine that runs your model server. See Install the agent.
2. Create an endpoint at engress.io, issue a connect token, and run engress-agent tunnel serve with your endpoint ID and local port. See Your first endpoint.
3. Use the public URL anywhere you would paste an API base URL. Requests hit the edge, traverse the outbound tunnel, and reach your local process.

Next steps

• Install the agent — macOS, Linux, or build from source.
• Connect tokens — what tokens are, how scopes work, and when to revoke.
• CLI reference — commands and flags for engress-agent.
• How it works — edge placement, TLS, and what happens when an edge drains.