Ollama Copilot

Ollama Copilot integrates local LLMs from Ollama directly into VS Code, providing AI-powered code completion and an interactive chat experience with your own locally-running models.

Changelog

Version 0.1.5

• 🚀 Improved inline suggestions with expanded context (up to 1000 lines)
• 🔄 Fixed Tab key acceptance for multi-line suggestions
• 🎯 Better code completion accuracy with enhanced context awareness
• 💡 Added support for more Ollama models including Qwen and Mixtral
• 🛠️ Improved error handling and connection stability
• 📝 Enhanced documentation with visual guides

Features

AI-Powered Code Completions

Get contextual code suggestions as you type, powered by your local Ollama models:

• Smart context awareness (up to 1000 lines of surrounding code)
• Multi-line code suggestions
• Language-specific completions
• Variable and function name awareness
• Tab completion support

Interactive Chat Interface

Engage with your code through:

• Dedicated sidebar chat panel
• Real-time streaming responses
• Context-aware code discussions
• File and workspace context integration

Privacy-Focused

• All processing happens locally through Ollama
• No data sent to external servers
• Complete control over your models and data

Customizable Configuration

• Choose from any installed Ollama model
• Configure API host settings
• Adjust workspace context settings

Prerequisites

1. Install Ollama on your system
2. Pull at least one model in Ollama (see model recommendations)
3. Make sure Ollama is running (ollama serve)

Quick Start

1. Install the extension from VS Code marketplace
2. Run Ollama in the background (ollama serve)
3. Select a default model when prompted
4. Start coding to see inline suggestions
5. Use the sidebar chat for more complex queries

Model Selection

Choose your model through:

1. Command Palette (Ctrl+Shift+P or Cmd+Shift+P)
2. Type "Ollama Copilot: Select Default Model"
3. Pick from your installed models

Recommended Models

For the best experience, we recommend:

Code Completion

• qwen:14b - Excellent for general code completion
• codellama:13b - Strong at understanding context
• deepseek-coder:6.7b - Fast and efficient
• phind-codellama:34b - Great for complex completions

Chat Interface

• mixtral:8x7b - Strong reasoning and explanation
• llama2:13b - Good balance of speed and capability
• neural-chat:7b - Fast responses for simple queries

Installing Models

# Install recommended models
ollama pull qwen:14b
ollama pull codellama:13b
ollama pull mixtral:8x7b

# List installed models
ollama list

Usage Tips

Code Completion

• Type normally and wait for suggestions
• Press Tab to accept full suggestions
• Use → (right arrow) to accept word by word
• Clear completion cache if suggestions seem stale

Chat Interface

• Click the Ollama icon in the sidebar
• Use @ to reference files
• Select code before asking questions
• Toggle workspace context for broader awareness

Commands

Access via Command Palette (Ctrl+Shift+P or Cmd+Shift+P):

• Ollama Copilot: Select Default Model - Change your model
• Ollama Copilot: Clear Completion Cache - Reset suggestions
• Ollama Copilot: Open Chat Panel - Open chat interface
• Ollama Copilot: Search Available Models - View installed models

Configuration

Settings available in VS Code:

• ollama.defaultModel: Your preferred model
• ollama.apiHost: Ollama API endpoint (default: http://localhost:11434)

Troubleshooting

No Suggestions

1. Verify Ollama is running (ollama serve)
2. Check model is selected (Command Palette > Select Default Model)
3. Clear completion cache
4. Ensure cursor is at a valid completion point

Performance Issues

1. Try a smaller model
2. Clear completion cache
3. Check system resources
4. Reduce context size if needed

Connection Issues

1. Confirm Ollama is running
2. Check ollama.apiHost setting
3. Verify port 11434 is accessible
4. Restart VS Code if needed

Contributing

We welcome contributions! Please check our GitHub repository for:

• Bug reports
• Feature requests
• Pull requests
• Documentation improvements

License

MIT License