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

# Cherry Studio

> Connect AIHubMix in the Cherry Studio desktop client to call Claude, GPT, Gemini and more, and enable Claude prompt caching to cut cost and latency.

[**Download Cherry Studio AI**](https://easys.run/cherry-studio/)

## General Usage

1. Open **Settings** from the bottom-left corner of the app.

   <img src="https://mintcdn.com/aihubmix/njYPCqTk0m3HBNaY/public/en/CS1.png?fit=max&auto=format&n=njYPCqTk0m3HBNaY&q=85&s=df290eeaf273e02bc81b1219ce2affbd" alt="Open Settings from the bottom-left corner of Cherry Studio" width="1356" height="835" data-path="public/en/CS1.png" />
2. In the **Model Providers** section, select **AiHubMix** and click the button in the top-right corner to enable it.
3. In the **API Key** field, enter [the Key from this site](https://aihubmix.com/token).\
   Leave the **API Base URL** unchanged.\
   **Note: If the connection check fails, try disabling your VPN.**

   <img src="https://mintcdn.com/aihubmix/njYPCqTk0m3HBNaY/public/en/CS2.png?fit=max&auto=format&n=njYPCqTk0m3HBNaY&q=85&s=d8940f62c356ed2ab9c5a5893286cbe9" alt="Select AiHubMix in Model Providers and enter the API Key" width="1358" height="837" data-path="public/en/CS2.png" />
4. Click **Add Model** at the bottom.\
   For the **Model ID**, go to the settings page on this website, select the model you want to use, and copy-paste its name.

   <img src="https://mintcdn.com/aihubmix/njYPCqTk0m3HBNaY/public/en/CS3.png?fit=max&auto=format&n=njYPCqTk0m3HBNaY&q=85&s=4108a6e7b9eb612f83456fa1039aa7b7" alt="Click Add Model and paste the model ID" width="1358" height="835" data-path="public/en/CS3.png" />

## FAQ

### The model indicates it cannot see images

Some newer models may not recognize images after they are uploaded.\
In this case, you need to manually enable the **Image** option on the settings page.

<img src="https://mintcdn.com/aihubmix/AfNgblL-wCNs82Ww/public/en/CS4.png?fit=max&auto=format&n=AfNgblL-wCNs82Ww&q=85&s=b6e45ddcbc89e351852a86badb39168b" alt="Enable the Image option on the settings page so the model can read images" width="483" height="281" data-path="public/en/CS4.png" />

<Tip>
  The `gemini-3-pro-image-preview` model cannot return images when **streaming output** is enabled.\
  Please disable streaming output when using this model.

  <img src="https://mintcdn.com/aihubmix/oCJ4o1SNFsJN6z-I/public/en/CS5.png?fit=max&auto=format&n=oCJ4o1SNFsJN6z-I&q=85&s=debdd87af29a3477568a471d20b51646" alt="Disable streaming output so images return correctly" width="2144" height="1554" data-path="public/en/CS5.png" />
</Tip>

### Calling Claude models returns `AI_APICallError`

Check whether the request URL shown in the error message is\
`https://aihubmix.com/anthropic/v1/messages`.

If so, go to **Settings** and change the **Anthropic API Host** to: `https://aihubmix.com`

<img src="https://mintcdn.com/aihubmix/G8QySdgS7UmbtB3g/public/en/CS6.png?fit=max&auto=format&n=G8QySdgS7UmbtB3g&q=85&s=8e67b948fc475bdcebbb3174a79d94e5" alt="Change the Anthropic API Host to aihubmix.com in Settings" width="1284" height="994" data-path="public/en/CS6.png" />

## Enabling Claude prompt caching

In the Aihubmix provider's "API Settings" (the gear to the right of the provider name), configure "Cache Token Threshold", "Cache System Message", and "Cache Last N Messages" to automatically inject `cache_control` for Claude models. For what each option means and cache-hit tips, see [Claude Prompt Caching](../api/Claude-Cache).

***

Last updated: 2026-06-01