Skip to content

README.md

Latest commit

 

History

History
159 lines (120 loc) · 4.87 KB

README.md

File metadata and controls

159 lines (120 loc) · 4.87 KB

WaddleAI for VS Code Copilot Chat

Integrate WaddleAI's enterprise AI proxy with VS Code Copilot Chat for enhanced AI-powered development with multi-LLM support, conversation memory, and security features.

Features

🚀 Multi-LLM Support: Access OpenAI GPT, Anthropic Claude, and Ollama models through a single interface
🧠 Conversation Memory: Context-aware responses that remember your development patterns
🔒 Enterprise Security: Built-in prompt injection protection and security scanning
📊 Token Management: Track usage across different models with normalized billing
⚡ Intelligent Routing: Automatic model selection based on cost, latency, and availability
🔧 Custom Configuration: Flexible endpoint configuration for on-premises deployments

Installation

  1. Install from VS Code Marketplace: Search for "WaddleAI" in the Extensions panel
  2. Get a WaddleAI API Key: Visit your WaddleAI management portal to generate an API key
  3. Configure the Extension: Use Ctrl+Shift+P → "WaddleAI: Set API Key"

Quick Start

1. Set Your API Key

Ctrl+Shift+P → WaddleAI: Set API Key

Enter your API key (format: wa-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx)

2. Select a Model

Ctrl+Shift+P → WaddleAI: Select Model

Choose from available models like GPT-4, Claude-3, or Llama2

3. Use in Copilot Chat

  1. Open Copilot Chat (Ctrl+Shift+I)
  2. Click the model selector dropdown
  3. Select "WaddleAI" as your provider
  4. Start chatting with AI assistance!

Configuration

Settings

Configure WaddleAI through VS Code settings (File → Preferences → Settings → Search "WaddleAI"):

{
  "waddleai.apiEndpoint": "http://localhost:8000",
  "waddleai.defaultModel": "gpt-4",
  "waddleai.enableMemory": true,
  "waddleai.enableSecurityScanning": true,
  "waddleai.maxTokens": 2048,
  "waddleai.temperature": 0.7
}

Available Models

  • GPT Models: gpt-4, gpt-3.5-turbo
  • Claude Models: claude-3-opus-20240229, claude-3-sonnet-20240229
  • Ollama Models: llama2, codellama (requires Ollama server)

Memory Enhancement

Enable conversation memory to get context-aware responses that remember:

  • Your coding preferences and patterns
  • Project-specific information
  • Previous conversations and solutions

Security Features

  • Prompt Injection Detection: Automatically scans for malicious prompts
  • Jailbreak Prevention: Blocks attempts to bypass AI safety guidelines
  • Data Protection: Prevents accidental exposure of sensitive information

Commands

Command Description
WaddleAI: Set API Key Securely store your WaddleAI API key
WaddleAI: Select Model Choose which AI model to use
WaddleAI: Test Connection Verify connection to WaddleAI server
WaddleAI: Show Token Usage View your current token usage and quotas
WaddleAI: Clear Memory Reset conversation memory

Advanced Usage

Custom Endpoints

For enterprise deployments, configure your own WaddleAI endpoint:

{
  "waddleai.apiEndpoint": "https://waddleai.yourcompany.com"
}

Code Context Integration

WaddleAI automatically includes:

  • Currently selected code in your questions
  • Active workspace information
  • File language context for better responses

Usage Monitoring

Track your AI usage with built-in analytics:

  • Token consumption per model
  • Daily and monthly quota tracking
  • Request history and patterns

Troubleshooting

Common Issues

❌ "Authentication failed"

  • Verify your API key is correct
  • Check that your WaddleAI account is active
  • Ensure proper permissions are granted

❌ "Connection failed"

  • Verify the WaddleAI server is running
  • Check your endpoint configuration
  • Ensure firewall/proxy settings allow access

❌ "Model not available"

  • Check which models are enabled in your WaddleAI configuration
  • Verify model permissions for your API key
  • Try selecting a different model

Debug Mode

Enable debug logging:

{
  "waddleai.debug": true
}

View logs in: View → Output → WaddleAI

Requirements

  • VS Code: Version 1.85.0 or higher
  • WaddleAI Server: Running proxy server with API access
  • API Key: Valid WaddleAI API key with appropriate permissions

Privacy & Security

  • API keys are stored securely in VS Code's secret storage
  • Communication with WaddleAI uses HTTPS encryption
  • No code or conversations are stored locally by default
  • Memory features can be disabled for maximum privacy

Support

License

MIT License - see LICENSE for details.

Developed by WaddleAI | Website | Enterprise