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

# Chatbox

> Configure AIOHub as an OpenAI-compatible model provider in Chatbox.

[Chatbox](https://chatboxai.app) is a cross-platform AI chat client with custom model providers. Use AIOHub through Chatbox's OpenAI-compatible provider mode.

Chatbox is especially useful for using the AIOHub API on mobile devices. If you want to chat with your own AIOHub API key on iOS, Android, or a tablet, Chatbox is usually more direct than developer-tool clients.

## Prerequisites

* Chatbox installed, or access to the Chatbox web or mobile app
* An [AIOHub API key](/en/account/create-token)
* A group assigned to the API key that can access the Chat Completions model you want to use
* A positive account balance

## Setup

<Steps>
  <Step title="Open model provider settings">
    Open Chatbox, click the settings entry in the sidebar, then open **Model Provider**.
  </Step>

  <Step title="Add a custom provider">
    Click **Add** or **Add Custom Provider** at the bottom of the provider list:

    | Field    | Value                 |
    | -------- | --------------------- |
    | Name     | AIOHub                |
    | API Mode | OpenAI API Compatible |
  </Step>

  <Step title="Enter API settings">
    Fill in the fields used by your Chatbox version:

    | Field    | Value                    |
    | -------- | ------------------------ |
    | API Key  | `sk-your-api-key`        |
    | API Host | `https://api.aiohub.org` |
    | API Path | `/v1/chat/completions`   |

    <Note>If your Chatbox version labels the field as Base URL and expects `/v1` there, use `https://api.aiohub.org/v1` and keep API Path as `/chat/completions`. Do not put `/v1` in both fields.</Note>
  </Step>

  <Step title="Add models">
    Add or enable the models you want to use:

    | Model                      | Good for                               |
    | -------------------------- | -------------------------------------- |
    | `gpt-4o`                   | Everyday chat, writing, translation    |
    | `gpt-4o-mini`              | Lightweight and lower-cost chat        |
    | `claude-sonnet-4-20250514` | Long-form analysis and complex writing |

    Available models depend on the group assigned to your API key. Use the console model list and `/v1/models` response as the source of truth.
  </Step>

  <Step title="Check the connection">
    Save the provider, then click **Check**. If the check succeeds, return to the chat page and select the configured model.
  </Step>

  <Step title="Start chatting">
    Create a new conversation and send a short message. You can then open the AIOHub console **Logs** page to review the request and cost.
  </Step>
</Steps>

## Recommended fields

| Chatbox field | AIOHub value                      | Notes                         |
| ------------- | --------------------------------- | ----------------------------- |
| API Mode      | OpenAI API Compatible             | Uses OpenAI Chat Completions  |
| API Host      | `https://api.aiohub.org`          | Use with the API Path below   |
| API Path      | `/v1/chat/completions`            | Chat Completions request path |
| API Key       | `sk-your-api-key`                 | Use the full API key value    |
| Model         | Any model visible to your API key | For example `gpt-4o`          |

## Optional settings

* If automatic chat title generation creates extra requests, disable it in Chatbox settings or use a cheaper helper model.
* Vision, tool use, and image features depend on the Chatbox version, model capabilities, and the group assigned to the API key.
* If you only need basic text chat, verify that works before enabling advanced features.

## Troubleshooting

<AccordionGroup>
  <Accordion title="Connection check fails">
    Confirm the API key starts with `sk-`, API Host includes `https://`, and API Path is `/v1/chat/completions`. If Host already includes `/v1`, use `/chat/completions` as the path.
  </Accordion>

  <Accordion title="The request returns 404 or repeats path segments">
    The usual cause is putting `/v1` in both API Host and API Path. The full AIOHub request URL should be `https://api.aiohub.org/v1/chat/completions`.
  </Accordion>

  <Accordion title="A model is unavailable">
    Check in the AIOHub console that the group assigned to the API key includes that model. You can also request `/v1/models` to see the models visible to the current API key.
  </Accordion>

  <Accordion title="Can Chatbox use Claude models through AIOHub?">
    Yes, when the model is available through the OpenAI Chat Completions-compatible path and visible to your API key. For native Claude Messages behavior, use a client with an Anthropic provider or Claude Cowork 3P.
  </Accordion>
</AccordionGroup>

## Official references

* [Chatbox model configuration](https://docs.chatboxai.app/en/guides/providers)
* [Chatbox OpenAI-compatible API guide](https://docs.chatboxai.app/en/chatbox-ai-premium/openai-compatible-api)