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

# Introduction

> Overview of DGrid's three API surfaces — the Model API, Management API Keys, and x402 pay-per-inference API — with authentication and base URLs.

The DGrid API Reference is organized into three API surfaces, each with its own authentication model and purpose:

<CardGroup cols={3}>
  <Card title="Model API" icon="brain" href="#model-api">
    OpenAI-, Claude-, and Gemini-compatible endpoints for chat, completions, embeddings, images, audio, and moderations.
  </Card>

  <Card title="Management API Keys" icon="key" href="/management-api-keys">
    Programmatic lifecycle management for Model API Keys — create, rotate, limit, disable, and revoke.
  </Card>

  <Card title="x402 API" icon="credit-card" href="/x402/overview">
    Pay-per-inference access using the x402 payment protocol, with no pre-funded account balance.
  </Card>
</CardGroup>

<Note>
  Every Model API reference page has an interactive playground: fill in your API key and parameters, then press **Send** to test the endpoint live against `https://api.dgrid.ai`.
</Note>

## Model API

The Model API gives you a single gateway to 200+ leading AI models. Every endpoint is compatible with an ecosystem standard — OpenAI, Claude, or Gemini — so your existing SDKs and tooling work unchanged.

For request-level cost reconciliation, store the `DGrid-Request-ID` response header and use [Get request billing details](/api-reference/usage-and-billing/get-request-billing-details) to retrieve token usage, pricing, and USD cost details.

### Base URL

```text theme={null}
https://api.dgrid.ai
```

### Authentication

All Model API endpoints authenticate with a bearer token:

```http theme={null}
Authorization: Bearer <DGRID_API_KEY>
```

The Claude-compatible `POST /v1/messages` endpoint additionally requires:

```http theme={null}
anthropic-version: 2023-06-01
```

Get your key by following the [key creation guide](https://blog.dgrid.ai/posts/2026-01-04/), or use the [Management API Keys](/management-api-keys) surface for programmatic key lifecycle management.

### Model naming

The `model` parameter uses the `provider/model-name` format — for example `openai/gpt-4o` — consistent across all DGrid-supported models. Call [List models](/api-reference/models/list-models) to discover the model IDs available to your key.

### Endpoint families

| Family          | Base path                                                                                     | Best for                                                                    |
| --------------- | --------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------- |
| Chat            | `/v1/chat/completions`, `/v1/responses`, `/v1/messages`, `/v1/models/{model}:generateContent` | Conversational apps, multimodal assistants, tool calling                    |
| Completions     | `/v1/completions`                                                                             | Prompt-only text generation and legacy integrations                         |
| Embeddings      | `/v1/embeddings`, `/v1/engines/{engine}/embeddings`                                           | Search, clustering, retrieval, ranking                                      |
| Images          | `/v1/images/generations`, `/v1/images/edits`                                                  | Image generation and editing across model formats                           |
| Audio           | `/v1/audio/speech`, `/v1/audio/transcriptions`, `/v1/audio/translations`                      | Speech synthesis, transcription, translation                                |
| Moderations     | `/v1/moderations`                                                                             | Content safety filtering and policy checks                                  |
| Models          | `/v1/models`                                                                                  | Discover available models and supported endpoint types                      |
| Usage & Billing | `/api/v1/model-router/billing-json`                                                           | Retrieve request-level token usage, pricing snapshots, and USD cost details |
| Realtime        | `/v1/realtime` (WebSocket)                                                                    | Low-latency streaming conversations                                         |

### Best practices

1. Store the DGrid API key on the server side and never expose it in public clients or repositories.
2. Prefer the OpenAI-compatible format when you want the broadest SDK compatibility with minimal migration effort.
3. Use provider-native formats only when you need provider-specific payload features such as Gemini multimodal parts or Claude tool payloads.
4. Keep request and response examples in sync with your target model family, because parameter names differ between OpenAI, Claude, and Gemini styles.

## Management API Keys

Management API Keys are administrative credentials for the full lifecycle of Model API Keys — creation, rotation, usage limits, disabling, and revocation. They are restricted to key management operations and cannot call Model API inference endpoints.

See [Management API Keys](/management-api-keys) for the API surfaces, authentication boundaries, and endpoint reference.

## x402 API

The x402 API is DGrid's pay-per-inference access mode. Instead of creating an account or pre-funding a balance, requests are authorized and settled per call using the x402 payment protocol — well suited to autonomous agents and on-demand integrations.

See [x402 API](/x402/overview) for the payment flow and [x402 API Reference](/x402/api-reference) for endpoint details.