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

# NexusAPI Documentation

> Get from creating an API key to your first NexusAPI request quickly.

<Info>
  NexusAPI provides a unified model gateway. Your token group, model availability, and pricing are determined by the **Model Marketplace** in the console.
</Info>

<Columns cols={3}>
  <Card title="Make your first request in 5 minutes" icon="rocket" href="/en/guide/quick-start" arrow>
    Create a key, choose a model, and send a test request.
  </Card>

  <Card title="Configure Claude Code / Codex" icon="terminal" href="/en/tools" arrow>
    Choose the setup guide for the client you use.
  </Card>

  <Card title="Confirm API and model capability" icon="braces" href="/en/developer/capability-matrix" arrow>
    Check that the model, token group, and protocol match before calling.
  </Card>
</Columns>

## The three things you need to connect

| Item     | Where to get it             | Purpose                                    |
| -------- | --------------------------- | ------------------------------------------ |
| Base URL | This documentation          | Sends requests to NexusAPI                 |
| API Key  | Console → Token Management  | Identifies the caller and available groups |
| Model ID | Console → Model Marketplace | Selects the model to call                  |

<Warning>
  Do not expose an API key in chats, screenshots, frontend repositories, or public configuration. If you suspect exposure, delete or reset the key in the console immediately.
</Warning>

## Start with your goal

<Columns cols={2}>
  <Card title="I am writing an application or script" icon="code-2" href="/en/guide/quick-start" arrow>
    Start with the OpenAI-compatible API.
  </Card>

  <Card title="I use Claude Code" icon="bot" href="/en/tools/claude-code" arrow>
    Configure the native Messages protocol.
  </Card>

  <Card title="I use Codex CLI" icon="terminal-square" href="/en/tools/codex-cli" arrow>
    Configure a custom provider and model.
  </Card>

  <Card title="A request failed" icon="circle-help" href="/en/faq/self-check" arrow>
    Follow the self-check flow and collect the required details.
  </Card>
</Columns>

## Include these details when requesting support

* Time of the issue and time zone
* Model ID, endpoint, and whether streaming was enabled
* Full error message and Request ID
* The smallest reproduction you have already tried

This prevents repeated follow-up questions and speeds up troubleshooting.