Z.ai Chat Provider for VS Code

Integrates Z.ai (智谱AI) models into VS Code Copilot Chat with advanced features including vision support, tool calling, and thinking process display.

Features

• Multiple Model Support

  • GLM-4.7: 200K context window, up to 128K output tokens
  • GLM-4.7 Flash: Faster variant with 131K max output tokens
  • GLM-5: 200K context window, up to 128K output tokens
  • GLM-4.6V: Vision model (internal only, not exposed to users)

• Advanced Capabilities

  • Tool calling support for VS Code chat participants
  • Streaming responses via Server-Sent Events (SSE)
  • Vision support via GLM-OCR and GLM-4.6V fallback
  • Thinking/reasoning process display (configurable)
  • Automatic image-to-text conversion for non-vision models

• Secure API Key Management

  • Stored securely in VS Code SecretStorage
  • Managed via Command Palette (Z.ai: Manage Z.ai Provider)

Installation

From Marketplace (Coming Soon)

code --install-extension Ryosuke-Asano.zai-vscode-chat

From Source

1. Clone the repository:

git clone https://github.com/Ryosuke-Asano/zai-provider-extension.git
cd zai-provider-extension

2. Install dependencies:

npm install

3. Compile the project:

npm run compile

4. Package the extension:

npm run package

5. Install the .vsix file:

code --install-extension zai-vscode-chat-*.vsix

Setup

1. Open VS Code
2. Open Command Palette (Cmd/Ctrl + Shift + P)
3. Run Z.ai: Manage Z.ai Provider
4. Enter your Z.ai API key

Get your API key from Z.ai Platform.

Usage

Once configured, select Z.ai as your chat provider in VS Code Copilot Chat:

• Open the Chat view (Cmd/Ctrl + Alt + I)
• Click the provider selector
• Choose a Z.ai model (GLM-4.7, GLM-4.7 Flash, or GLM-5)
  • Note: GLM-4.6V is used internally for image processing and is not selectable

Configuration

Supported Models

User-Selectable Models

Internal Models (Not Exposed)

MCP Integration

This extension integrates with Z.ai's MCP (Model Context Protocol) servers:

• web-search-prime: Web search capabilities
• web-reader: URL to text/markdown conversion
• zread: GitHub repository file reading
• vision-mcp: Image analysis

Development

See CONTRIBUTING.md for detailed development guidelines.

Quick Start

# Install dependencies
npm install

# Watch for changes
npm run watch

# Run tests
npm test

# Lint code
npm run lint

# Format code
npm run format

Project Structure

src/
├── extension.ts    # Extension entry point, activation
├── provider.ts     # Main chat provider implementation
├── types.ts        # Type definitions and model configuration
├── mcp.ts          # MCP client for tool integration
└── utils.ts        # Utility functions for message/tool conversion

Requirements

• VS Code 1.104.0 or later
• Node.js 20 or later (for development)
• Z.ai API key

Troubleshooting

API Key Issues

If you see authentication errors:

1. Run Z.ai: Manage Z.ai Provider
2. Verify your API key is correct
3. Ensure your API key has active credits

Vision Not Working

For non-vision models (GLM-4.7, GLM-4.7 Flash):

• Images are automatically converted to text descriptions using GLM-OCR MCP
• If GLM-OCR fails, the extension internally uses GLM-4.6V for image analysis
• GLM-4.6V is not selectable by users—it is only used as an internal fallback

Large Context Errors

If you encounter token limit errors:

• Reduce the amount of code/context in your message
• The extension enforces model-specific context limits

Changelog

See CHANGELOG.md for version history.

License

MIT © 2025 Ryosuke Asano

License

Links

• Repository
• Issue Tracker
• Z.ai Platform