Skip to content

Zed Integration

Zed is a high-performance, collaborative code editor designed for speed and efficiency. It includes built-in AI assistance features that can be customized to use different language models.

  • Collaborative Editing: Real-time collaboration with peers
  • AI Assistant Integration: Built-in AI code assistance
  • High Performance: Rust-based editor optimized for speed
  • Multiple Cursor Support: Efficient multi-line editing
  • Project-Wide Search: Fast code navigation and search
  • Collaborative software development
  • Pair programming sessions with AI assistance
  • Remote team coding with shared context
  • High-performance code editing with AI suggestions
  1. Install Zed

    Follow the Zed installation instructions

  2. Open Zed preferences

    Open Zed preferences CMD/CTRL + Shift + P

    Zed Command Palette

  3. Open settings file

    Navigate to zed:open settings to open the settings.json file.

  4. Add custom LLM provider

    Add a custom LLM provider with these settings:

    {
    "language_models": {
    "openai": {
    "api_url": "https://api.relax.ai/v1",
    "available_models": [
    {
    "name": "DeepSeek-R1-0528",
    "display_name": "Deepseek R1 0528",
    "max_tokens": 131072,
    "max_output_tokens": null
    }
    ],
    "version": "1"
    }
    },
    }
  5. Set default assistant model (optional)

    If you would like to use the custom model as default assistant, set the following parameters in the same settings.json file:

    "assistant": {
    "default_model": {
    "provider": "openai",
    "model": "DeepSeek-R1-0528"
    },
    }

    Zed Customization

  6. Configure API key

    Make sure to set your RELAX_API_KEY to the environment variable OPENAI_API_KEY or configure the key under OpenAI section in the chat configuration.

  7. Save and restart

    Save changes and restart Zed.

    Zed with Custom Model

To use the AI assistant in Zed with your custom model:

  1. Press Cmd+I (Mac) or Ctrl+I (Windows/Linux) to open the assistant
  2. Type your query or request
  3. The assistant will respond using your configured relaxAI model