Connect Your Self‑Hosted LLM to Volcengine Edge Gateway in 4 Simple Steps

Compared with third‑party provider models, self‑deployed models offer data sovereignty, high customization, and stability, making them suitable for privacy‑sensitive, vertical‑scenario, and long‑term controllable applications.

This guide shows how to call a self‑deployed model through Volcengine Edge Large Model Gateway without writing any code, using only four simple steps.

Step 1: Add the self‑deployed model

Log in to the Edge Large Model Gateway console, navigate to Model Configuration Management → Large Model Management, select the “Self‑Deployed Model” tab and add a model. Provide model name, call name (identifier), model series, provider (self‑deployed only), format (OpenAI), and model type.

Step 2: Create a calling channel

Provide the model’s API key and choose a proxy mode (local service proxy, LAN service proxy, or domain proxy) to enable inbound‑outbound tunneling for models that only support intranet access.

Local service proxy – for models running on the same machine; set local port and run the proxy tool after channel creation.

LAN service proxy – for models on a LAN server; set IP address and port, then run the proxy tool on any LAN server.

Domain proxy – for models exposed via a public domain; set an OpenAI‑compatible base URL.

Step 3: Bind the channel to a gateway access key

Create or edit a gateway access key and bind the previously created calling channel so that API requests can target the self‑deployed model.

Step 4: Use the gateway access key

Integrate the provided sample code with the bound access key; when the request is sent, the gateway forwards it to the specified self‑deployed model.

The Edge Large Model Gateway also provides authentication, rate limiting, security, and monitoring to manage the full lifecycle of large‑model applications.