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

# API addresses and authentication

> Choose the SDK Base URL, full endpoint, and authentication header for each protocol.

Send API requests to `api.apione.ai`. The console at `apione.ai/en/app` manages accounts and keys; it is not an API Base URL.

![APIONE API address, protocols, and key management](https://mintcdn.com/apione/pAFvFGkbmcRZirTZ/images/ui/api-keys-overview.png?fit=max\&auto=format\&n=pAFvFGkbmcRZirTZ\&q=85\&s=286a882ab53c81ad3bd16d56c6487dc9)

The API Key page shows APIONE Base URL, supported protocols, discount groups, and model prices.

## Address reference

| Protocol or purpose | SDK Base URL | Full HTTP endpoint |
| - | - | - |
| [Chat Completions](https://apione.ai/docs/en/models/chat.md) | `https://api.apione.ai/v1` | `POST https://api.apione.ai/v1/chat/completions` |
| [OpenAI Responses](https://apione.ai/docs/en/api/responses.md) | `https://api.apione.ai/v1` | `POST https://api.apione.ai/v1/responses` |
| [Anthropic Messages](https://apione.ai/docs/en/api/anthropic.md) | `https://api.apione.ai` | `POST https://api.apione.ai/v1/messages` |
| [Gemini](https://apione.ai/docs/en/api/gemini.md) | `https://api.apione.ai/v1beta` | `POST https://api.apione.ai/v1beta/models/{model}:generateContent` |
| [Model list](https://apione.ai/docs/en/models/catalog.md) | `https://api.apione.ai/v1` | `GET https://api.apione.ai/v1/models` |
| [Model parameters](https://apione.ai/docs/en/models/parameters.md) | `https://api.apione.ai/v1` | `GET https://api.apione.ai/v1/models/{model}/params` |

SDKs usually append the endpoint; HTTP requests need the full address. Some clients append the version path themselves. For example, Gemini CLI uses the root URL; follow its [guide](https://apione.ai/docs/en/guides/gemini-cli.md). The final path should contain `/v1` or `/v1beta` only once.

## Request headers

| Interface | Authentication header |
| - | - |
| OpenAI-compatible APIs | `Authorization: Bearer YOUR_APIONE_KEY` |
| Anthropic Messages | `x-api-key: YOUR_APIONE_KEY` |
| Gemini | `x-goog-api-key: YOUR_APIONE_KEY` |

Use one authentication credential per request. JSON requests also need `Content-Type: application/json`. Anthropic Messages examples send `anthropic-version: 2023-06-01`; see the [integration guide](https://apione.ai/docs/en/api/anthropic.md).

## Track a request

Save the request-identifier headers actually returned (`X-Request-ID`, `X-Maxx-Trace-Id`, or `X-Maxx-Request-ID`), retaining each name and value, HTTP status, and request time for console lookup or support. Remove API Keys and sensitive business content from logs and reports.

## Models and protocols

An APIONE API Key authenticates requests to the protocols provided by the platform. Confirm protocol support in the model details, then use the [parameters API](https://apione.ai/docs/en/models/catalog.md) to inspect the model's declared extra parameters and allowed values. An endpoint's availability does not establish support for every parameter on every model.

See the [endpoint overview](https://apione.ai/docs/en/api/endpoints.md) for image generation, image editing, and video APIs. Follow the selected model's request requirements. Image input in Chat and image generation are different capabilities.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.