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

# Agent Auto-Setup

> Universal setup guide for any AI agent or OpenAI-compatible client. Just provide your API key.

## Universal Config

Every AI agent that supports OpenAI-compatible APIs works with Kyma. Here's everything you need:

```yaml theme={null}
base_url: https://kymaapi.com/v1
api_key: YOUR_KYMA_KEY  # starts with ky-
```

That's it. Two values. Everything else is auto-discovered via `/v1/models`.

## Auto-Detection

Kyma API keys start with `ky-`. Configure:

| Setting       | Value                    |
| ------------- | ------------------------ |
| Base URL      | `https://kymaapi.com/v1` |
| API Key       | The `ky-` key            |
| Format        | OpenAI-compatible        |
| Streaming     | Supported                |
| Default model | `qwen-3.6-plus`          |

## Quick Verify

```bash theme={null}
curl https://kymaapi.com/v1/models \
  -H "Authorization: Bearer YOUR_KEY"
```

Returns all available models. Pick any one.

## Agent-Specific Configs

<AccordionGroup>
  <Accordion title="Claude Code / Claude CLI">
    ```json theme={null}
    {
      "apiProvider": "openai-compatible",
      "baseUrl": "https://kymaapi.com/v1",
      "apiKey": "ky-your-api-key",
      "model": "qwen-3.6-plus"
    }
    ```
  </Accordion>

  <Accordion title="Cursor IDE">
    Settings → OpenAI API:

    * **API Key**: `ky-your-api-key`
    * **Base URL**: `https://kymaapi.com/v1`
    * **Model**: `qwen-3-32b` (best for code)
  </Accordion>

  <Accordion title="Continue.dev">
    `~/.continue/config.json`:

    ```json theme={null}
    {
      "models": [{
        "provider": "openai",
        "title": "Kyma - Qwen 3.6 Plus",
        "model": "qwen-3.6-plus",
        "apiBase": "https://kymaapi.com/v1",
        "apiKey": "ky-your-api-key"
      }]
    }
    ```
  </Accordion>

  <Accordion title="OpenAI Python SDK">
    ```python theme={null}
    from openai import OpenAI

    client = OpenAI(
        base_url="https://kymaapi.com/v1",
        api_key="ky-your-api-key"
    )

    response = client.chat.completions.create(
        model="qwen-3.6-plus",
        messages=[{"role": "user", "content": "Hello!"}]
    )
    ```
  </Accordion>

  <Accordion title="OpenAI Node.js SDK">
    ```typescript theme={null}
    import OpenAI from 'openai';

    const client = new OpenAI({
      baseURL: 'https://kymaapi.com/v1',
      apiKey: 'ky-your-api-key',
    });

    const response = await client.chat.completions.create({
      model: 'qwen-3.6-plus',
      messages: [{ role: 'user', content: 'Hello!' }],
    });
    ```
  </Accordion>

  <Accordion title="LangChain">
    ```python theme={null}
    from langchain_openai import ChatOpenAI

    llm = ChatOpenAI(
        base_url="https://kymaapi.com/v1",
        api_key="ky-your-api-key",
        model="qwen-3.6-plus"
    )
    ```
  </Accordion>

  <Accordion title="Any OpenAI-compatible client">
    ```yaml theme={null}
    base_url: https://kymaapi.com/v1
    api_key: ky-your-api-key
    model: qwen-3.6-plus
    ```

    Works with: Hermes, OpenClaw, Aider, Cline, Windsurf, LibreChat, anything OpenAI-compatible.
  </Accordion>
</AccordionGroup>

## Recommended Models

| Use Case            | Model              | Speed  |
| ------------------- | ------------------ | ------ |
| General / reasoning | `qwen-3.6-plus`    | Medium |
| Coding              | `qwen-3-coder`     | Medium |
| Long context        | `gemini-2.5-flash` | ⚡ Fast |
| Deep reasoning      | `deepseek-r1`      | Slow   |
| Agentic tasks       | `kimi-k2.6`        | Medium |
| Fastest response    | `qwen-3-32b`       | ⚡ Fast |

## MCP Integration

Connect your AI agent to Kyma docs for auto-discovery:

```json theme={null}
{
  "mcpServers": {
    "kyma-docs": {
      "url": "https://docs.kymaapi.com/mcp"
    }
  }
}
```

Your agent can then search Kyma documentation and auto-configure itself.