Copilot ++

Overview#

Copilot ++ (also known as better-copilot-chat) is a VS Code extension developed by OEvortex, designed to extend GitHub Copilot Chat's model support capabilities. Through a unified provider registry system, users can seamlessly switch between 20+ AI providers directly within the Copilot Chat interface.

Core Capabilities#

Multi AI Provider Support#

Major providers include:

• Antigravity (Google Cloud Code): Gemini 3 Pro, Gemini 3.1 Pro, Gemini 3 Flash
• Codex (OpenAI): GPT-5.2 Codex, GPT-5.3 Codex
• ZhipuAI (GLM Coding Plan): GLM-4.5, GLM-4.6, GLM-4.7, GLM-5, GLM-4.7-Flash
• MiniMax (Coding Plan): MiniMax-M2.5, MiniMax-M2.1
• MoonshotAI (Kimi For Coding): Kimi-K2-Thinking, Kimi-K2-0905-Preview
• DeepSeek: DeepSeek-V3.2, DeepSeek-V3.2 Reasoner

Other providers: Chutes, OpenCode, Blackbox, DeepInfra, Kilo AI, Zenmux, Lightning AI, Hugging Face, Mistral AI, NVIDIA NIM, Ollama Cloud, Qwen CLI, Gemini CLI, and Compatible Provider for custom OpenAI/Anthropic compatible models.

Multi-Account Management#

• Support for unlimited accounts per provider
• Quick account switching (Ctrl+Shift+Q / Cmd+Shift+Q)
• Account status display in status bar
• Secure credential storage using VS Code Secret Storage

Load Balancing & Auto-Switching#

• Automatic request distribution across accounts
• Auto-switch on rate limits or quota exhaustion
• Smart retry with exponential backoff
• Real-time quota monitoring and usage statistics

OAuth Authentication#

• Antigravity: Google OAuth login
• Codex: OpenAI OAuth login
• Gemini CLI: Google OAuth (via CLI)
• Qwen CLI: Alibaba OAuth (via CLI)

Web Search Integration#

• #zhipuWebSearch: Multi-engine search (Sogou, Quark, Standard)
• #minimaxWebSearch: Coding Plan web search
• #googleWebSearch: Grounded Google Search with citations

Advanced Code Completion#

• FIM (Fill In the Middle): Context-aware intelligent code completion
• NES (Next Edit Suggestions): Predictive edit suggestions
• Support for manual (Alt+/) and auto-trigger modes

Edit Tool Modes#

• claude mode: ReplaceString - efficient single replacement (default)
• gpt-5 mode: ApplyPatch - batch editing, complex refactoring
• none mode: Direct file editing - fallback for edge cases

Requirements#

• VS Code >= 1.104.0
• Node.js >= 20.0.0 (for source build)
• npm >= 9.0.0 (for source build)
• GitHub Copilot Chat extension (required dependency)

Installation#

VS Code Marketplace (Recommended)#

In VS Code, press Ctrl+P and paste:
ext install OEvortex.better-copilot-chat

Install from .vsix file#

code --install-extension better-copilot-chat-0.2.5.vsix

Build from Source#

git clone https://github.com/OEvortex/better-copilot-chat.git
cd better-copilot-chat
npm install
npm run compile
npm run package

Quick Start#

1. Configure Provider: Cmd/Ctrl+Shift+P → Select e.g. Copilot ++: ZhipuAI Configuration Wizard
2. Select Model: Click model dropdown in Copilot Chat → Choose configured model
3. Enable Load Balancing: Cmd/Ctrl+Shift+P → Copilot ++: Settings → Toggle Load Balance

Global Configuration Example#

{
    "chp.temperature": 0.1,
    "chp.topP": 1,
    "chp.maxTokens": 8192,
    "chp.editToolMode": "claude",
    "chp.rememberLastModel": true
}

Keyboard Shortcuts#

Project Info#

• Publisher: OEvortex
• Current Version: 0.2.5
• Primary Languages: TypeScript (94.0%), JavaScript (4.7%), CSS (1.3%)
• License: MIT License