# Connect through the proxy

Use `http://localhost:18082` as your app's LLM base URL.

## Main endpoint

```text
http://localhost:18082
```

## Supported client style

The Community proxy is intended to work with:

- OpenAI-compatible clients
- Anthropic-compatible clients
- Gemini-compatible clients

## Typical OpenAI-compatible setup

```bash
OPENAI_BASE_URL=http://localhost:18082
```

Some clients call the same setting `OPENAI_API_BASE`.

## Login key vs provider key

There are two different keys in the Community flow:

- the bootstrap tenant API key is for logging into the dashboard on `http://localhost:18080`
- your provider API key stays in your AI app or IDE for model requests

## Quick check

If your app still works after switching its LLM base URL to `http://localhost:18082`, Cencurity is in the traffic path.
