Zed Editor

Configure Zed's built-in AI features to use Chuizi.AI for inline code generation and chat with any model.

Prerequisites

• Zed editor installed
• A Chuizi.AI account with an API key (starts with ck-)

Configuration

1. Open Zed settings: Cmd+, (macOS) or via menu Zed > Settings
2. Add the following to your settings.json:

{
  "language_models": {
    "openai": {
      "api_url": "https://api.chuizi.ai/v1",
      "api_key": "ck-your-key-here",
      "available_models": [
        {
          "name": "anthropic/claude-sonnet-4-6",
          "display_name": "Claude Sonnet 4.6",
          "max_tokens": 16000
        },
        {
          "name": "openai/gpt-4.1",
          "display_name": "GPT-4.1",
          "max_tokens": 32768
        },
        {
          "name": "deepseek/deepseek-chat",
          "display_name": "DeepSeek Chat",
          "max_tokens": 8192
        }
      ]
    }
  }
}

You can add any Chuizi.AI-supported model to the available_models array.

Usage

Once configured, use the AI features through:

• Inline assist: Select code and press Cmd+Enter, then type your instruction
• AI chat panel: Open with Cmd+?
• Code completion: Triggered automatically while writing code

Verify the Configuration

1. Open any code file
2. Select a block of code and press Cmd+Enter
3. Type "explain this code"
4. If you receive an AI response, the configuration is working

Common Issues

Models Not Showing in Settings

Make sure you have correctly populated the available_models array. Zed requires you to manually declare available models and will not automatically fetch them from the API.

AI Features Not Responding

Check that api_url is set to https://api.chuizi.ai/v1 (note: v1 without a trailing slash). Confirm your api_key starts with ck-.

Setting max_tokens

max_tokens controls the maximum tokens per AI response. Recommended values: 4096 for code completions, 16000 for conversations.

Next Steps

• Choose a Model — compare coding-focused models
• Pricing — model pricing details
• First Request — verify your setup works end-to-end