Models
Custom endpoints
Point Goatfied at your self-hosted vLLM, Ollama, or any OpenAI-compatible URL.
Configure
Settings → Models → Custom provider. Provide a base URL, optional auth header, and a model name. Goatfied probes the endpoint and surfaces the model in the picker.
{
"name": "internal-llama",
"baseUrl": "https://llm.internal.example.com/v1",
"model": "llama-3.3-70b",
"headers": { "x-team": "platform" }
}