> ## Documentation Index
> Fetch the complete documentation index at: https://docs.gpuhub.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Exposing Multiple Services

> Currently, the custom services provided by GPUhub only support exposing a single port. If you need to expose multiple services (HTTP services) from different containers, you can use a proxy to implement routing and forwarding. For example, you can use Nginx for forwarding. Here, we provide a more lightweight approach.

## Scope of the Lightweight Proxy

The lightweight proxy solution (route\_path-based forwarding) is primarily designed for backend API and HTTP request forwarding.

It does not guarantee full compatibility with **complex Web UI applications, such as ComfyUI, Gradio, or other SPA-based interfaces**. These applications typically rely on static asset routing, frontend path resolution, and WebSocket connections, which may not function correctly under simple path-based forwarding.

If you need to expose a complete Web UI application, we recommend deploying a standard reverse proxy (e.g., Nginx) inside the container to handle routing and asset forwarding properly.

## Route-Based Forwarding with proxy\_in\_instance

Assuming you have two services running inside your container, listening on ports `2000` and `3000` respectively, here is an example using Flask:

```python python theme={null}
# Simulate a service listening on port 2000
from flask import Flask
app = Flask(__name__)

@app.route('/v1/hello')
def hello():
    return 'I`m port 2000'

if __name__ == '__main__':
    app.run(port=2000)


# Simulate a service listening on port 3000
from flask import Flask
app = Flask(__name__)

@app.route('/v2/hello')
def hello():
    return 'I`m port 3000'

if __name__ == '__main__':
    app.run(port=3000)
```

Below, we will start a routing and forwarding service on port `6006`. If the route matches `/v1/*`, it will forward the request to the service listening on port `2000`. If the route matches `/v2/*`, it will forward the request to the service listening on port `3000`.
D
ownload the proxy service program:

First, download the proxy service program using the following command:

```bash theme={null}
wget https://autodl-public.ks3-cn-beijing.ksyuncs.com/tool/api-proxy/proxy_in_instance
```

Create the `config.yaml` file:

In the same directory as the downloaded proxy service program, create a file named config.yaml with the following content:

```yaml yaml theme={null}
proxies:
  - host_and_port: http://127.0.0.1:2000  # Address of the service to proxy
    route_path: /v1/*                    # Route path to forward to this address

  - host_and_port: http://127.0.0.1:3000  # Multiple entries can be set to forward to different hosts
    route_path: /v2/*
```

Start the proxy service:

```bash bash theme={null}
# If the proxy program is not downloaded, download it using the command above.
wget https://autodl-public.ks3-cn-beijing.ksyuncs.com/tool/api-proxy/proxy_in_instance
```

```yaml yaml theme={null}
# If the proxy program does not have executable permissions, add the permissions
chmod +x proxy_in_instance
```

```bash bash theme={null}
# Start the proxy service:
./proxy_in_instance
```