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

# Portrix API Reference: Endpoints and Authentication

> Complete reference for the Portrix REST API. OpenAI-compatible endpoints for chat, completions, embeddings, and model listing at api.portrix.ai/v1.

The Portrix API is a REST API that follows the OpenAI API specification. Any tool, SDK, or code that works with OpenAI works with Portrix with minimal changes — simply swap the base URL and your API key to gain access to 400+ models from every major provider through a single, unified endpoint.

## Base URL

All API endpoints are relative to the following base URL:

```
https://api.portrix.ai/v1
```

Every request you make should target a path under this base URL. For example, to call the chat completions endpoint you send a request to `https://api.portrix.ai/v1/chat/completions`.

## Authentication

Portrix authenticates requests using Bearer tokens in the `Authorization` header. You can generate an API key from your dashboard at [app.portrix.ai](https://app.portrix.ai).

```http theme={null}
Authorization: Bearer <your-api-key>
```

See the [Authentication guide](/api-reference/authentication) for detailed setup instructions, SDK examples, and security best practices.

## Endpoints

<CardGroup cols={2}>
  <Card title="Chat Completions" icon="messages" href="/api-reference/chat-completions">
    **POST /v1/chat/completions**

    Generate chat responses from any model using a conversation history. The primary endpoint for most use cases.
  </Card>

  <Card title="Text Completions" icon="text" href="/api-reference/completions">
    **POST /v1/completions**

    Legacy prompt-based text completion for workflows that require a raw continuation rather than a chat turn.
  </Card>

  <Card title="Embeddings" icon="vector-square" href="/api-reference/embeddings">
    **POST /v1/embeddings**

    Convert text into numeric vector representations for semantic search, clustering, and retrieval-augmented generation.
  </Card>

  <Card title="Models" icon="list" href="/api-reference/models">
    **GET /v1/models**

    Retrieve the full list of AI models available through Portrix, including provider metadata and capability flags.
  </Card>
</CardGroup>

## Request Format

All requests must use a JSON body with the `Content-Type: application/json` header set. All responses are also returned as JSON. For streaming endpoints, responses use `text/event-stream` (Server-Sent Events).

```http theme={null}
Content-Type: application/json
Authorization: Bearer <your-api-key>
```

## Model IDs

Portrix uses a `{provider}/{model}` naming convention for all model IDs. This tells the gateway which upstream provider to route your request to. For example:

| Provider  | Example Model ID              |
| --------- | ----------------------------- |
| OpenAI    | `openai/gpt-4o`               |
| Anthropic | `anthropic/claude-3-5-sonnet` |
| Google    | `google/gemini-2.0-flash`     |
| Meta      | `meta/llama-3.1-70b-instruct` |

Pass the full `provider/model` string in the `model` field of any request body.

## Portrix-Specific Headers

In addition to the standard OpenAI-compatible headers, Portrix supports several request headers that control routing, fallback, and search behavior. You can attach these to any request.

| Header                 | Type     | Description                                                                                                                               |
| ---------------------- | -------- | ----------------------------------------------------------------------------------------------------------------------------------------- |
| `x-portrix-model`      | `string` | Override the model used for a request without changing the request body `model` field. Useful for middleware that cannot modify the body. |
| `x-portrix-fallback`   | `string` | A comma-separated ordered list of fallback model IDs to try if the primary model is unavailable.                                          |
| `x-portrix-route`      | `string` | Force routing to a specific provider region or deployment tier.                                                                           |
| `x-portrix-search`     | `string` | Enable or configure web-search augmentation for supported models (`"true"` to enable).                                                    |
| `x-portrix-request-id` | `string` | **Response header.** A unique ID for the request, useful for debugging and support tickets.                                               |

## Rate Limits

Portrix enforces rate limits on a per-key basis. When you exceed a limit, the API returns a `429 Too Many Requests` response with a `Retry-After` header indicating when you can retry. See the [Rate Limits guide](/configuration/rate-limits) for information on default limits, how to check your current usage, and how to request an increase.

## SDKs

You can use Portrix with any OpenAI-compatible SDK by setting the `base_url` and `api_key`. No custom SDK required.

<CardGroup cols={2}>
  <Card title="Python SDK" icon="python" href="/integrations/python">
    Use the official `openai` Python package pointed at the Portrix base URL.
  </Card>

  <Card title="TypeScript / Node.js SDK" icon="js" href="/integrations/javascript">
    Use the `openai` npm package with a custom `baseURL` option.
  </Card>

  <Card title="REST / cURL" icon="terminal" href="/integrations/openai-sdk">
    Call the API directly with any HTTP client — no SDK required.
  </Card>

  <Card title="LangChain & LlamaIndex" icon="link" href="/integrations/langchain">
    Drop Portrix in as the LLM provider in popular orchestration frameworks.
  </Card>
</CardGroup>