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

# Cherry Studio Integration

> Use the platform with Cherry Studio desktop AI client

## Overview

[Cherry Studio](https://cherry-ai.com/) is a cross-platform desktop AI client that supports multiple LLM providers. Configure the platform as a custom provider to access a broad range of models with TEE protection through Cherry Studio's beautiful interface.

<Info>
  Cherry Studio provides a ChatGPT-like interface for Windows, macOS, and Linux with advanced features like multi-model chat and local LLM support.
</Info>

## Installation

Download Cherry Studio from:

* **macOS**: [cherry-studio.dmg](https://github.com/CherryHQ/cherry-studio/releases)
* **Windows**: [cherry-studio.exe](https://github.com/CherryHQ/cherry-studio/releases)
* **Linux**: [cherry-studio.AppImage](https://github.com/CherryHQ/cherry-studio/releases)

## Adding the Platform as Custom Provider

### Step 1: Open Settings

1. Launch Cherry Studio
2. Click **Settings** in the left navigation bar
3. Select the **Model Services** tab

### Step 2: Add Custom Provider

1. Click the **+ Add** button
2. Fill in the configuration:

<Steps>
  <Step title="Provider Name">
    Enter: `RedPill AI`
  </Step>

  <Step title="Provider Type">
    Select: `OpenAI`
  </Step>

  <Step title="API Address">
    Enter: `https://api.redpill.ai/v1`

    <Note>
      Cherry Studio automatically appends `/chat/completions`. Just use the base URL ending with `/v1`
    </Note>
  </Step>

  <Step title="API Key">
    Enter your API key: `YOUR_API_KEY`

    Click the **Test** button to verify the connection
  </Step>
</Steps>

### Step 3: Add Models

Click **+ Add** in the models section and add your preferred models:

**Popular Models:**

```
z-ai/glm-5.1
z-ai/glm-5
qwen/qwen3.5-27b
openai/gpt-5
anthropic/claude-sonnet-4.5
anthropic/claude-opus-4.1
deepseek/deepseek-chat-v3.1
google/gemini-2.5-pro
```

<Info>
  Click the **Manage** button to auto-fetch all available models, then select which ones to add.
</Info>

## Complete Configuration Example

```json theme={null}
{
  "provider_name": "RedPill AI",
  "provider_type": "OpenAI",
  "api_base": "https://api.redpill.ai/v1",
  "api_key": "YOUR_API_KEY",
  "models": [
    "z-ai/glm-5.1",
    "z-ai/glm-5",
    "qwen/qwen3.5-27b",
    "openai/gpt-5",
    "anthropic/claude-sonnet-4.5",
    "deepseek/deepseek-chat-v3.1"
  ]
}
```

## Using the Models in Cherry Studio

### 1. Start a Chat

1. Click the **+** button to create a new chat
2. Select a model from the dropdown
3. Start chatting!

### 2. Multi-Model Comparison

Cherry Studio allows comparing responses from multiple models:

1. Create a new chat

2. Enable **Multi-Model** mode

3. Select multiple models:
   * `openai/gpt-5` for general tasks
   * `anthropic/claude-sonnet-4.5` for reasoning
   * `deepseek/deepseek-chat-v3.1` for coding

4. Send one message and get responses from all models simultaneously

### 3. Model-Specific Settings

Configure per-model settings:

* **Temperature**: Control randomness (0.0 - 2.0)
* **Max Tokens**: Limit response length
* **Top P**: Nucleus sampling parameter
* **Presence Penalty**: Reduce repetition
* **Frequency Penalty**: Encourage diversity

## Advanced Configuration

### Custom API Routes

If the platform uses non-standard routes, you can specify the full path:

**Standard route (auto-appended):**

```
https://api.redpill.ai/v1
→ Becomes: https://api.redpill.ai/v1/chat/completions
```

**Custom route (use # to prevent appending):**

```
https://api.redpill.ai/v2/custom/completions#
→ Uses exactly: https://api.redpill.ai/v2/custom/completions
```

### Environment-Specific Configurations

Create multiple configurations:

1. **Production**
   * Name: `RedPill Production`
   * API: `https://api.redpill.ai/v1`
   * Models: Production models only

2. **Testing**
   * Name: `RedPill Testing`
   * API: `https://api-test.redpill.ai/v1`
   * Models: Experimental models

### Vision Models Configuration

Add vision-capable models for image analysis:

```
openai/gpt-5
anthropic/claude-sonnet-4.5
google/gemini-2.5-pro
qwen/qwen3-vl-30b-a3b-instruct
```

Then upload images in your chats for analysis.

## Features Available with the Platform

<CardGroup cols={2}>
  <Card title="Multi-Model Chat" icon="comments">
    Compare responses from multiple models simultaneously
  </Card>

  <Card title="Streaming" icon="stream">
    Real-time streaming responses from all models
  </Card>

  <Card title="Image Analysis" icon="image">
    Upload images with vision-capable models
  </Card>

  <Card title="Chat History" icon="clock">
    Save and search conversation history
  </Card>

  <Card title="Code Highlighting" icon="code">
    Automatic syntax highlighting for code
  </Card>

  <Card title="Markdown Support" icon="markdown">
    Rich markdown rendering
  </Card>

  <Card title="Export Chats" icon="download">
    Export conversations as Markdown or JSON
  </Card>

  <Card title="Dark Mode" icon="moon">
    Beautiful dark theme support
  </Card>
</CardGroup>

## Recommended Model Setup

### For General Use

Add these models for everyday tasks:

```
openai/gpt-5              # Best for general chat
anthropic/claude-sonnet-4.5  # Best for analysis
deepseek/deepseek-chat-v3.1     # Best for coding
```

### For Privacy-Sensitive Work

Use GPU TEE confidential models:

```
z-ai/glm-5.1
z-ai/glm-5
qwen/qwen3.5-27b
```

Requests are served through the attested TEE gateway, which does not retain request bodies. Only confidential (`is_tee`) models attest the upstream provider.

### For Multilingual Support

```
openai/gpt-5
google/gemini-2.5-pro
anthropic/claude-sonnet-4.5
```

## Troubleshooting

<AccordionGroup>
  <Accordion title="Connection Test Fails">
    **Solutions:**

    * Verify API key is correct
    * Check API base URL ends with `/v1`
    * Ensure internet connection is active
    * Try removing and re-adding the provider
  </Accordion>

  <Accordion title="Models Not Loading">
    **Solutions:**

    * Click **Manage** → **Refresh** to fetch models
    * Manually add model IDs with provider prefix
    * Verify API key has proper permissions
  </Accordion>

  <Accordion title="Responses Are Slow">
    **Solutions:**

    * Check your internet connection
    * Try a faster model (e.g., `google/gemini-2.5-flash`)
    * Reduce `max_tokens` in model settings
  </Accordion>

  <Accordion title="Images Not Working">
    **Solutions:**

    * Ensure you're using a vision-capable model
    * Models supporting images: GPT-5, Claude Sonnet 4.5, Gemini 2.5
    * Check image file size (under 20MB recommended)
  </Accordion>
</AccordionGroup>

## Cherry Studio Features

### 1. Assistant Presets

Create custom assistants with:

* Pre-defined system prompts
* Specific model preferences
* Temperature and other settings

**Example: Code Review Assistant**

```
Model: deepseek/deepseek-chat-v3.1
Temperature: 0.3
System Prompt: "You are a senior software engineer. Review code for bugs, performance issues, and best practices."
```

### 2. Quick Commands

Use `/` commands in chat:

* `/clear` - Clear chat history
* `/export` - Export conversation
* `/model` - Switch model mid-conversation

### 3. File Upload

Upload files for analysis:

* **Images** - Vision models can analyze
* **Code files** - AI can review and explain
* **Documents** - Summarize and extract info

### 4. Search History

Search through all past conversations:

* Keyword search
* Filter by model
* Filter by date range

## Best Practices

<AccordionGroup>
  <Accordion title="Model Selection">
    * **GPU TEE models**: Best for sensitive data and proprietary work
    * **GLM 5.1**: Best for confidential reasoning
    * **GLM-5**: Best for confidential systems engineering
    * **GPT-5**: Best for general conversation
    * **Claude 4.5**: Best for reasoning and analysis
    * **DeepSeek**: Best for coding tasks
    * **Gemini Flash**: Best for speed
  </Accordion>

  <Accordion title="Multi-Model Usage">
    Use multi-model mode to:

    * Compare different approaches
    * Validate AI responses
    * Find the best model for your task
    * Get diverse perspectives
  </Accordion>

  <Accordion title="Privacy">
    * Requests are served through the attested TEE gateway, which does not retain request bodies
    * Only confidential (`is_tee`) models attest the upstream provider
    * Use GPU TEE models for end-to-end confidential AI
    * Cherry Studio stores chats locally
    * Export important conversations for backup
  </Accordion>

  <Accordion title="Cost Management">
    * Monitor usage in the dashboard
    * Use faster/cheaper models for simple tasks
    * Set max\_tokens to control costs
    * Export chats to avoid re-processing
  </Accordion>
</AccordionGroup>

## Comparison with Other Clients

| Feature          | Cherry Studio             | ChatGPT Web   | Claude.ai     |
| ---------------- | ------------------------- | ------------- | ------------- |
| Multi-Provider   | ✅ A broad range of models | ❌ OpenAI only | ❌ Claude only |
| Local Storage    | ✅ Private                 | ❌ Cloud       | ❌ Cloud       |
| Multi-Model Chat | ✅ Yes                     | ❌ No          | ❌ No          |
| TEE Protection   | ✅ With the platform       | ❌ No          | ❌ No          |
| Offline Mode     | ✅ With local LLMs         | ❌ No          | ❌ No          |
| Open Source      | ✅ Yes                     | ❌ No          | ❌ No          |

## Alternative Providers in Cherry Studio

You can add multiple providers simultaneously:

1. **This platform**: a broad range of models with TEE protection
2. **OpenAI Direct**: direct OpenAI access
3. **Local LLMs**: Ollama, LM Studio
4. **Other Services**: any API

Switch between providers based on your needs.

## Resources

* [Cherry Studio GitHub](https://github.com/CherryHQ/cherry-studio)
* [Cherry Studio Documentation](https://docs.cherry-ai.com/)
* [Models list](/api-reference/models)
* [Discord](https://discord.gg/P2ukR4Z5ps)

## Next Steps

<CardGroup cols={2}>
  <Card title="Supported Models" icon="list" href="/api-reference/models">
    Browse all available models
  </Card>

  <Card title="Vision Guide" icon="image" href="/guides/vision">
    Learn about image analysis
  </Card>

  <Card title="Privacy Features" icon="shield" href="/confidential-ai/how-it-works">
    Understand TEE protection
  </Card>

  <Card title="Confidential Models" icon="lock" href="/confidential-ai/confidential-models">
    Explore confidential AI models
  </Card>
</CardGroup>