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

# Introduction

> OpenAI-compatible inference API on the Gonka network — fast start, no vendor lock-in.

Dahl gives you an **OpenAI-compatible API** for open models running on decentralized GPU infrastructure. Use the same endpoints and request shape you already know (`/v1/models`, `/v1/chat/completions`), with a Bearer API key from your [account](https://inference.dahl.global/account) or `POST /tokens`.

**Base URL:** `https://inference.dahl.global/v1`

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/quickstart">
    Get an API key and send your first chat completion in under five minutes.
  </Card>

  <Card title="Try chat" icon="message" href="https://inference.dahl.global/chat">
    Test models in the browser before wiring them into your app.
  </Card>

  <Card title="Create account" icon="user" href="https://inference.dahl.global/account">
    Username + fingerprint — first API key issued immediately.
  </Card>

  <Card title="Models" icon="cube" href="/models">
    See which model IDs are available and which one to use by default.
  </Card>
</CardGroup>

## How it works

1. **Get a key** — create an [account](https://inference.dahl.global/account) (recommended) or call `POST https://inference.dahl.global/tokens`.
2. **List models** — `GET /v1/models` (no API key required).
3. **Call the API** — `POST /v1/chat/completions` with `Authorization: Bearer` and your key.

The recommended default model is **MiniMax M2.7** (`MiniMaxAI/MiniMax-M2.7`). Other models are listed on the network; always pick IDs from `GET /v1/models` rather than hardcoding stale values.

See [Authentication](/authentication) for accounts, fingerprints, key claim/revoke, and balances.