Getting Started

Migrate from OpenAI

Draven AI mirrors the OpenAI API's request and response shapes. Any client already built for OpenAI works by changing two things.

Setting	Value
Base URL	`https://api.dravenai.lat/v1`
API Key	Your `sk-drv-...` key from the console

Official SDKs

client = OpenAI(
    base_url="https://api.dravenai.lat/v1",  # was: default OpenAI URL
    api_key="sk-drv-your-api-key",           # was: sk-... (OpenAI key)
)

const client = new OpenAI({
  baseURL: "https://api.dravenai.lat/v1", // was: default OpenAI URL
  apiKey: "sk-drv-your-api-key",          // was: sk-... (OpenAI key)
});

Environment variable swap

Many OpenAI-compatible tools (LiteLLM, Open WebUI, LibreChat, Vercel AI SDK) read the base URL and key from environment variables. Setting these is often the only change needed:

.env

OPENAI_BASE_URL=https://api.dravenai.lat/v1
OPENAI_API_KEY=sk-drv-...

What to check after switching

Model name — use an id from Models (e.g. your-model-id) instead of an OpenAI model id like gpt-4o.
Rate limits — Draven AI's limits are per-user and listed in Rate Limits; they aren't the same numbers as your OpenAI plan.
Unsupported OpenAI parameters — the request body accepts model, messages, stream, temperature, top_p, and max_tokens (see Chat Completions). Other OpenAI-only parameters are currently ignored rather than erroring.

Errors keep the same shape as OpenAI's ({ "error": { "message", "type", "param", "code" } }), so existing error-handling and retry logic doesn't need to change. See Errors.