ChatGLM Router for GitHub Copilot Chat

A VS Code extension forked from Hugging Face's huggingface-vscode-chat project, modified to integrate ChatGLM (with Coding and General endpoints) into GitHub Copilot Chat.

Compatibility

Due to limited testing resources, compatibility testing may not be exhaustive. Please report any issues on GitHub or in the comments section, and I will address them promptly.

AI Assistance Statement

The development of this plugin utilized ChatGLM Coding to complete most of the API adaptation work.

Demo

Usage Demo

Quick Start

Install the ChatGLM Router extension (search for "ChatGLM Router" in VS Code extensions)
Open VS Code Copilot Chat interface (Ctrl/Cmd + Shift + A)
Click the model picker and click "Manage Models..."
Find "ChatGLM Router" and click "Manage ChatGLM Router"
Select the provider you want to manage and enter its API Key. For ChatGLM, you can get one from https://open.bigmodel.cn/
Choose the models you want to add to the model picker

📋 Roadmap

Recently Implemented:

[x] Multiple Custom Providers - Support for other OpenAI-compatible APIs (DeepSeek, OpenAI, Azure OpenAI, local LLMs, etc.)
[x] Real-time Token Usage - Status bar shows weekly/monthly token usage
[x] Usage Statistics - Detailed request and token statistics with weekly/monthly histograms

Planned features for future releases:

[ ] Token Alias Support - Use custom API endpoints while maintaining unified token billing
[ ] Streaming Token Usage - Real-time token count display during chat
[ ] Usage Cost Estimation - Calculate API costs based on token consumption
[ ] Export Usage Reports - Export statistics to CSV/JSON for further analysis
[ ] Usage Alerts - Notify when approaching API quota limits
[ ] Multi-language Model Names - Support for models with non-English identifiers
[ ] Model Caching - Cache model lists for faster loading

Have a suggestion? Feel free to open an issue on GitHub!

💖 Support This Project

Enjoy using this extension? Your support helps me continue developing and maintaining it!

🚀 Get GLM Coding Premium at a Discount - Use my referral link to get a special deal on GLM Coding subscription:

20+ Programming Tools: Seamlessly supports Claude Code, Cline, and more
Enhanced Coding Power: Supercharge your development workflow
Limited Time Offer: Exclusive discount for new users

Get GLM Coding Premium →

GLM Coding Premium

By subscribing through this link, you get a premium experience while supporting the development of this extension at no extra cost. Thank you for your support! 🙏

Available Models

ChatGLM Coding (Default)

Optimized for code generation and programming tasks
Endpoint: https://open.bigmodel.cn/api/coding/paas/v4

ChatGLM General (Optional)

For general chat and non-coding tasks
Endpoint: https://open.bigmodel.cn/api/paas/v4/
Enable in settings if needed (disabled by default)
Same models available, optimized for conversational AI

Custom Providers (Now Supported)

This extension now supports adding custom API providers for any OpenAI-compatible API.

Supported Provider Presets

We have built-in presets for popular providers to make configuration easy:

OpenAI - GPT-4, GPT-3.5, and more
DeepSeek - DeepSeek-V3, DeepSeek-Coder
Anthropic - Claude series models
Azure OpenAI - Azure-hosted OpenAI services
Zhipu AI (GLM) - Zhipu GLM series models
Moonshot AI (Kimi) - Moonshot Kimi series models
Baichuan - Baichuan series models
MiniMax - MiniMax series models
Ollama (Local) - Local Ollama service
LM Studio (Local) - Local LM Studio service
Together AI - Together AI API
Groq - Groq ultra-fast inference API
Cerebras - Cerebras ultra-fast inference API
OpenRouter - OpenRouter API (access to multiple models)

How to Add Custom Providers

Method 1: Add from Presets (Recommended)

Press Ctrl/Cmd + Shift + P to open the command palette
Run "ChatGLM Router: Add Custom API Provider"
Select "Select from Presets"
Choose a preset from the list (e.g., DeepSeek, OpenAI, etc.)
Confirm the provider ID (can be customized)
Run "ChatGLM Router: Manage Custom Providers" to configure API Key

Method 2: Manual Configuration

Press Ctrl/Cmd + Shift + P to open the command palette
Run "ChatGLM Router: Add Custom API Provider"
Select "Manual Configuration"
Fill in the following information:
- Provider ID: Unique identifier (e.g., my-openai)
- Display Name: Name shown in UI (e.g., My OpenAI)
- API Base URL: API endpoint (e.g., https://api.openai.com/v1)
- Model Family: Model family identifier (e.g., openai)
- Supports Tool Calling: Whether Function Calling is supported
Run "ChatGLM Router: Manage Custom Providers" to configure API Key

Manage Custom Providers

Run "ChatGLM Router: Manage Custom Providers" to:

Add new custom providers
Configure API Key: Configure or update API keys for existing providers
View details: View provider configuration information
Delete provider: Remove unused custom providers

Configure in VS Code Settings

You can also configure directly in settings.json:

{
  "chatglmRouter.customProviders": [
    {
      "id": "deepseek",
      "name": "DeepSeek",
      "baseUrl": "https://api.deepseek.com/v1",
      "family": "deepseek",
      "supportsTools": true
    },
    {
      "id": "my-openai",
      "name": "My OpenAI",
      "baseUrl": "https://api.openai.com/v1",
      "family": "openai",
      "supportsTools": true
    }
  ]
}

After configuration, run "ChatGLM Router: Manage Custom Providers" to configure API keys for each provider.

Configuration

API Key

ChatGLM API Key

Configure your ChatGLM API key via the command palette:

Press Ctrl/Cmd + Shift + P
Run "ChatGLM Router: Manage"
Select "ChatGLM (Coding & General)"
Enter your API key from https://open.bigmodel.cn/

Custom Provider API Key

Configure API keys for custom providers:

Press Ctrl/Cmd + Shift + P
Run "ChatGLM Router: Manage Custom Providers"
Select the provider you want to configure
Select "Configure API Key"
Enter the API key for that provider
The extension will automatically verify the API key and display the number of available models

Clear API Key

Clear ChatGLM API Key

Run "ChatGLM Router: Clear ChatGLM API Key" to delete the stored ChatGLM API key

Clear Custom Provider API Key

Run "ChatGLM Router: Manage Custom Providers"
Select the provider you want to clear
Select "Configure API Key"
Clear the input field and confirm

Model Selection

Models are now displayed in the "Provider: Model Name" format:

ChatGLM Built-in Providers

ChatGLM Coding: glm-4.7 - ChatGLM Coding endpoint (default, recommended for VS Code)
ChatGLM Coding: glm-4-air - ChatGLM Coding lightweight model
ChatGLM General: glm-4.7 - ChatGLM General endpoint
ChatGLM General: glm-4-plus - ChatGLM General enhanced model

Custom Provider Examples

DeepSeek: deepseek-chat - DeepSeek-V3 non-reasoning mode
DeepSeek: deepseek-reasoner - DeepSeek-V3 reasoning mode
OpenAI: gpt-4 - OpenAI GPT-4
My OpenAI: gpt-4 - Custom-configured OpenAI

Notes:

If a provider doesn't have an API key configured, the model picker will show "Provider Name: Provider Name (API key not configured)"
Selecting an unconfigured model will prompt you to enter an API key
The same model (e.g., glm-4.7) can appear from different providers and will be displayed separately

Settings

Configure in VS Code Settings under chatglmRouter:

Setting	Type	Default	Description
`defaultProvider`	`string`	`chatglm-coding`	Default provider to use
`customProviders`	`object[]`	`[]`	Custom provider list (see below)
`statistics.enabled`	`boolean`	`true`	Enable usage statistics tracking
`statistics.statusBar.enabled`	`boolean`	`true`	Show statistics in status bar
`statistics.modelTooltip.enabled`	`boolean`	`true`	Show usage in model tooltips

Custom Provider Configuration

The customProviders setting supports the following fields:

Field	Type	Required	Description
`id`	string	✅	Unique provider identifier (e.g., `deepseek`)
`name`	string	✅	Display name (e.g., `DeepSeek`)
`baseUrl`	string	✅	API base URL (e.g., `https://api.deepseek.com/v1`)
`family`	string	✅	Model family identifier (e.g., `deepseek`)
`supportsTools`	boolean	❌	Whether tool calling is supported (default: true)
`defaultMaxTokens`	number	❌	Default max output tokens (default: 8192)
`defaultContextLength`	number	❌	Default context length (default: 128000)

Example Configuration:

{
  "chatglmRouter.customProviders": [
    {
      "id": "deepseek",
      "name": "DeepSeek",
      "baseUrl": "https://api.deepseek.com/v1",
      "family": "deepseek",
      "supportsTools": true,
      "defaultMaxTokens": 8192,
      "defaultContextLength": 128000
    }
  ]
}

For detailed statistics settings, see Statistics Settings below.

Usage Statistics

Track your API usage with built-in statistics:

Visualized Statistics

Run "ChatGLM Router: Show Usage Statistics" and select "显示可视化统计" to open a webview with charts.
View token usage per model for each provider.

Real-time Status Bar

Weekly and monthly token usage displayed in VS Code status bar
Auto-updates after each conversation request
Hover to see detailed statistics
Click to view full statistics

Hover over models in the picker to see historical usage
Shows total tokens, request count, and last used time
Helps you track which models you use most

View Statistics

Run "ChatGLM Router: Show Usage Statistics" command
View total requests and tokens per provider
See detailed per-model usage
Refresh statistics with confirmation feedback

Reset Statistics

Run "ChatGLM Router: Reset Usage Statistics" command
Clears all stored usage data

Statistics in Output

Run "ChatGLM Router: Show Statistics in Output" command
Displays detailed statistics in an output channel

Note: Statistics are stored locally in VS Code's global state and are estimates (4 chars ≈ 1 token).

Statistics Settings {#statistics-settings}

Configure in VS Code Settings under chatglmRouter.statistics:

Setting	Options	Default	Description
`statusBar.enabled`	boolean	true	Show statistics in status bar
`statusBar.displayMode`	normal, compact, minimal	normal	Status bar display mode
`statusBar.timeRange`	week, month, both	both	Time range to display
`statusBar.showRequestCount`	boolean	true	Show request count in status bar
`modelTooltip.enabled`	boolean	true	Show usage in model tooltips

Development

git clone https://github.com/OrientLuna/ChatGLM-vscode-chat
cd ChatGLM-vscode-chat
npm install
npm run compile

Press F5 to launch an Extension Development Host for testing.

Common Scripts

Build: npm run compile
Watch: npm run watch
Lint: npm run lint
Format: npm run format
Test: npm run test
Package: npm run package (generates .vsix file)

Architecture

Multi-Provider Design: Supports ChatGLM Coding, ChatGLM General, and custom providers
Provider Preset System: Built-in presets for popular providers with quick configuration
Provider Registry: Built-in providers configured in src/config.ts
Custom Providers: Add via VS Code settings or command palette
Statistics Tracking: Usage data tracked in src/statistics.ts
API-First Model List: Automatically fetches latest models from provider APIs
Streaming Response: SSE-like streaming with tool call support
Internationalization: Multi-language support (English and Chinese)

Troubleshooting

Models not appearing

ChatGLM models not appearing

Check that your ChatGLM API key is configured correctly
Run "ChatGLM Router: Manage" to verify API key
Check VS Code developer console for errors (Help → Toggle Developer Tools)

Custom provider models not appearing

Check provider configuration: Ensure the provider is correctly added to customProviders settings
Configure API Key: Run "ChatGLM Router: Manage Custom Providers" to configure API key
Verify API connection: When configuring API key, the extension will automatically verify and display available model count
Check API format: Ensure the API endpoint is OpenAI-compatible (supports /models endpoint)
Check console: Look for error messages in the developer console

API Errors

ChatGLM API errors

Verify your API key has the required permissions
Check that the selected model is available on your chosen endpoint
Ensure you have sufficient API credits/quotas

Custom provider API errors

Check API Key: Ensure the API key is correct and valid
Check endpoint URL: Verify the baseUrl is configured correctly (usually needs /v1 suffix)
Check model format: Ensure the API returns model list in OpenAI format
View error details: The model picker will display specific error messages
Check network connection: Ensure you can reach the API endpoint

Custom providers not working after adding

Reload window: Configuration requires running "Developer: Reload Window"
Check configuration: Ensure the configuration format in settings.json is correct
Check provider ID: Ensure the id field is unique and doesn't conflict with built-in providers

Local providers (Ollama, LM Studio) connection issues

Ensure service is running: Make sure Ollama or LM Studio is running
Check port: Verify the port is configured correctly (Ollama default 11434, LM Studio default 1234)
Check URL: Local addresses are typically http://localhost:11434/v1 or http://localhost:1234/v1

ChatGLM Coding vs General

Use ChatGLM Coding for code-related tasks (recommended for VS Code)
Use ChatGLM General for conversational AI and non-coding tasks
Enable ChatGLM General in settings: chatglmRouter.enabledProviders → add chatglm-general

Requirements

VS Code 1.104.0 or higher
ChatGLM API key from https://open.bigmodel.cn/

License

Support

Report issues: GitHub Issues
ChatGLM Documentation: https://open.bigmodel.cn/

ChatGLM Router for GitHub Copilot Chat

Orient Luna

ChatGLM Router for GitHub Copilot Chat

Compatibility

AI Assistance Statement

Demo

Quick Start

📋 Roadmap

💖 Support This Project

Available Models

ChatGLM Coding (Default)

ChatGLM General (Optional)

Custom Providers (Now Supported)

Supported Provider Presets

How to Add Custom Providers

Manage Custom Providers

Configure in VS Code Settings

Configuration

API Key

ChatGLM API Key

Custom Provider API Key

Clear API Key

Clear ChatGLM API Key

Clear Custom Provider API Key

Model Selection

ChatGLM Built-in Providers

Custom Provider Examples

Settings

Custom Provider Configuration

Usage Statistics

Visualized Statistics

Real-time Status Bar

Model Usage in Tooltip

View Statistics

Reset Statistics

Statistics in Output

Statistics Settings {#statistics-settings}

Development

Common Scripts

Architecture

Troubleshooting

Models not appearing

ChatGLM models not appearing

Custom provider models not appearing

API Errors

ChatGLM API errors

Custom provider API errors

Custom providers not working after adding

Local providers (Ollama, LM Studio) connection issues

ChatGLM Coding vs General

Requirements

License

Support