🦙 Ollama Agent - Local AI Coding Assistant

Transform your locally-running Ollama models into an intelligent coding agent with capabilities rivaling GitHub Copilot and Cursor—all running 100% locally with complete privacy. No cloud, no subscriptions, no data leaks.

🎉 New in v1.1: AI Completions, Quick Edit (Cmd+K), Code Review, Undo/Redo, Snippets Library, Prompt Templates, Theme Toggle, Favorites, and more!

✨ Key Features

🚀 Copilot-Like Features

• ✍️ AI Code Completions: Ghost text suggestions as you type (like GitHub Copilot)
• ⚡ Quick Edit (Cmd+K): Inline code editing with natural language instructions
• 🎯 Quick Actions Bar: One-click access to Explain, Fix, Optimize, Test, and more
• 📋 Prompt Templates Library: 10 pre-built templates for common tasks
• 💾 Code Snippets Library: Reusable code snippets for React, Node.js, Python, TypeScript
• ⏪ Undo/Redo System: Full history of AI edits with one-click undo (Ctrl+Alt+Z)
• 🔍 Code Review Mode: Comprehensive AI-powered code reviews

🎨 Enhanced UI/UX

• 🌓 Light/Dark Theme: Toggle between themes with sun/moon icon
• ⭐ Thread Favorites: Star important conversations and filter favorites
• 💬 Message Reactions: React to AI responses with 👍 👎 ❤️ 🎉
• ⌨️ Keyboard Shortcuts Panel: View all shortcuts in one place
• ✨ Smooth Animations: Professional fade-in, slide, and hover effects
• 📊 Performance Metrics: Real-time token count, response time, tokens/sec display

🔧 Advanced Context Gathering

• @terminal: Include terminal output in your prompts
• @diagnostics: Add compiler errors and warnings
• @workspace: Include project file structure
• @git-diff: Add uncommitted changes context
• @filename: Attach specific files to messages
• 🔍 History Search: Instantly find past conversations

🤖 Intelligent Agent Mode

• Autonomous Task Execution: Agent can read files, search codebases, create/edit files, and run tasks
• Multi-Step Reasoning: Up to 8-step problem-solving with context awareness
• Smart Tool Usage: Automatically explores your project structure and applies changes
• Read & Agent Modes: Choose between safe read-only or full agent capabilities

💬 Interactive Chat Interface

• Modern WebView UI: Clean, responsive chat interface with streaming responses
• Multi-Model Support: Chat with multiple Ollama models simultaneously
• Context-Aware: Automatically includes selected code, open files, and project structure
• Conversation History: Keep track of all interactions within sessions

📝 Code Generation & Editing

• Inline Code Generation: Generate code directly at cursor position
• Smart Refactoring: Simplify, optimize, or add types to selected code
• Function Extraction: Automatically extract functions with AI assistance
• Code Translation: Convert code between languages
• Multi-File Editing: Process multiple @mentioned files in one operation
• Direct Apply: Changes applied instantly with visual feedback

📚 Documentation & Analysis

• README Generator: Dedicated mode to generate comprehensive project documentation
• Docstring Completion: AI-generated documentation for functions and classes
• Code Explanation: Understand complex code with AI explanations
• PR Summaries: Generate pull request descriptions and review comments
• Code Review: Get detailed analysis of code quality, security, performance, and bugs

🧪 Testing & Quality

• Unit Test Generation: Create comprehensive test suites automatically
• Problem Analysis: AI-powered debugging and error resolution
• Code Review: Get suggestions for improvements and best practices

🛠️ Developer Tools

• Project Indexer: Fast semantic search across your entire codebase
• Commit Message Generation: Smart git commit messages from staged changes
• Shell Command Explanation: Understand complex terminal commands
• File Creation with AI: Generate boilerplate files with context

🎨 Customization

• Configurable Settings: Host, port, system prompts, and behavior modes
• Keyboard Shortcuts: Quick access to all commands
• Context Menu Integration: Right-click access in editor
• Status Bar Integration: Quick model switching and status updates

📦 Installation

Prerequisites

1. Install Ollama: Download from ollama.ai
2. Pull Models: Choose your preferred models

   ollama pull llama3.1:8b
   ollama pull mistral
   ollama pull codellama
   ollama pull deepseek-coder
   

3. Start Ollama: Ensure it's running on localhost:11434

   ollama serve
   

Install Extension

From VS Code Marketplace (Coming Soon):

1. Open VS Code
2. Go to Extensions (Ctrl+Shift+X / Cmd+Shift+X)
3. Search for "Ollama Agent"
4. Click Install

From VSIX (Manual):

1. Download the .vsix file from GitHub Releases
2. Open VS Code
3. Run command: Extensions: Install from VSIX...
4. Select the downloaded file

🚀 Quick Start

1. Open Chat Panel

• Click the Ollama icon in the Activity Bar (left sidebar)
• Or run command: Ollama Agent: Open Chat (Ctrl+Alt+K)
• The chat panel will open with Quick Actions bar

2. Try Quick Actions (NEW!)

• Click ⚡ Quick Actions to expand the action bar
• Choose from 8 one-click actions: Explain, Fix, Optimize, Tests, Document, Refactor, Debug, Review
• Actions automatically populate the input with slash commands

3. Use AI Completions (NEW!)

• Enable in Settings: ollamaAgent.enableInlineCompletions
• Start typing code and see ghost text suggestions appear
• Press Tab to accept, Esc to dismiss
• Toggle on/off with status bar icon

4. Quick Edit with Cmd+K (NEW!)

• Select code and press Ctrl+K (or Cmd+K on Mac)
• Type instruction: "add error handling" or "optimize this"
• Changes apply instantly with visual feedback
• Use Ctrl+Alt+Z to undo any AI edit

5. Select Agent Mode

• Read Mode (Safe): AI can only read and suggest changes
• Agent Mode (Powerful): AI can automatically create/edit files

3. Start Chatting

• Type your question or request
• Use @filename to attach files: @src/app.ts explain this
• Use @terminal, @diagnostics, @workspace, @git-diff for context
• Select one or more models to query
• Get streaming responses with real-time metrics

4. Explore New Features

• 📋 Templates Button: Access 10 prompt templates for common tasks
• 💾 Snippets Button: Copy ready-to-use code snippets
• ⌨️ Keyboard Shortcuts: View all shortcuts in modal
• 🌓 Theme Toggle: Switch between light/dark themes
• ⭐ Favorites: Star important conversations
• 🔍 Search: Find past conversations instantly

5. Use Context Menu Commands

• Select code → Right-click → Ollama Agent options
• Ask about selection, generate code, refactor, explain, and more

⚙️ Configuration

Settings

Access via File > Preferences > Settings → Search "Ollama Agent"

Example Configuration

{
  "ollamaAgent.host": "localhost",
  "ollamaAgent.port": 11434,
  "ollamaAgent.mode": "agent",
  "ollamaAgent.systemPrompt": "You are an expert TypeScript developer focused on clean code and best practices."
}

📖 Usage Examples

Generate a Function

1. Place cursor where you want the function
2. Run: Ollama Agent: Generate Code Here
3. Describe what you want: "Create a function to validate email addresses"
4. Review and insert the generated code

Refactor Code

1. Select the code to refactor
2. Right-click → Ollama Agent: Brush – Simplify
3. AI will suggest cleaner, more readable code
4. Review changes in diff view and apply

Create Unit Tests

1. Select a function or class
2. Run: Ollama Agent: Generate Unit Tests
3. AI generates comprehensive test suite
4. Tests appear in new file or clipboard

Generate README

1. Run: Ollama Agent: README Generator
2. Choose style (Technical/Product/Library)
3. Choose placement (GitHub/Marketplace/Website)
4. Add custom notes
5. Enable deep scan for comprehensive analysis
6. AI generates complete README with all sections

Debug with AI

1. Click on a problem in Problems panel
2. Run: Ollama Agent: Fix This Problem
3. AI analyzes error and suggests fixes
4. Apply fix with one click

🎯 Command Reference

New Commands (v1.1)

• Ollama Agent: Quick Edit (Ctrl+K / Cmd+K) - Inline code editing
• Ollama Agent: Code Review - Comprehensive AI code review
• Ollama Agent: Undo AI Edit (Ctrl+Alt+Z) - Undo last AI change
• Ollama Agent: Redo AI Edit (Ctrl+Alt+Shift+Z) - Redo AI change
• Ollama Agent: Show Edit History - View all AI edits
• Ollama Agent: Clear Edit History - Reset edit history
• Ollama Agent: Toggle Completions - Enable/disable ghost text

Chat & Interaction

• Ollama Agent: Open Chat - Open main chat panel
• Ollama Agent: README Generator - Generate project documentation
• Ollama Agent: Ask About Selection - Query AI about selected code

Code Generation

• Ollama Agent: Generate Code Here - Insert AI-generated code
• Ollama Agent: Generate Unit Tests - Create test suite
• Ollama Agent: Generate Commit Message - Smart git messages
• Ollama Agent: Generate README - Auto-generate documentation

Code Editing

• Ollama Agent: Edit Selection by Instruction - Guided refactoring
• Ollama Agent: Brush – Simplify - Simplify code
• Ollama Agent: Brush – Add Types - Add TypeScript types
• Ollama Agent: Brush – Optimize - Performance optimization
• Ollama Agent: Extract Function - Extract selected code

Analysis & Explanation

• Ollama Agent: Explain Selection - Understand code
• Ollama Agent: Analyze Problems - Debug assistance
• Ollama Agent: Explain Command - Shell command help
• Ollama Agent: Show Index Stats - View codebase stats

Project Management

• Ollama Agent: PR Summary - Generate PR descriptions
• Ollama Agent: PR Review Comments - AI code review
• Ollama Agent: Rebuild Project Index - Refresh semantic search
• Ollama Agent: Add File (AI) - Generate boilerplate files

Translation

• Ollama Agent: Translate Selection - Convert between languages

⌨️ Keyboard Shortcuts

🔧 Troubleshooting

Ollama Not Connecting

Problem: "Failed to connect to Ollama"

Solutions:

1. Verify Ollama is running: ollama serve
2. Check settings: host=localhost, port=11434
3. Test connection: curl http://localhost:11434/api/tags
4. Restart VS Code

No Models Available

Problem: "No models found"

Solutions:

1. Pull a model: ollama pull llama3.1
2. List models: ollama list
3. Refresh extension: Run Developer: Reload Window

Slow Response Times

Problem: AI responses are very slow

Solutions:

1. Use smaller models (e.g., llama3.1:8b instead of 70b)
2. Reduce context: Disable deep scan in README generator
3. Close other resource-intensive applications
4. Consider using GPU acceleration with Ollama

Agent Mode Not Working

Problem: Agent mode doesn't apply changes

Solutions:

1. Check mode setting: ollamaAgent.mode = agent
2. Verify file permissions (read/write access)
3. Check workspace trust: File → Trust Workspace
4. Review error messages in Output panel

Extension Not Loading

Problem: Extension doesn't activate

Solutions:

1. Check VS Code version (minimum 1.105.0)
2. View logs: Output → Ollama Agent
3. Disable conflicting extensions
4. Reinstall extension

🛠️ Development

Setup

# Clone repository
git clone https://github.com/IamNishant51/Ollama-Agent-vs-code-extension.git
cd Ollama-Agent-vs-code-extension

# Install dependencies
npm install

# Build webview
npm run build:webview

# Compile TypeScript
npm run compile

# Watch mode (auto-recompile)
npm run watch

Publishing

See PUBLISHING.md for a step-by-step guide to package and publish this extension to the VS Code Marketplace.

Testing

# Run linter
npm run lint

# Run tests
npm test

Debug

1. Open project in VS Code
2. Press F5 to launch Extension Development Host
3. Test features in new window
4. View logs in Debug Console

🤝 Contributing

Contributions are welcome! Please read our Contributing Guide for details.

Areas for Contribution

• Add support for more AI providers
• Improve agent reasoning capabilities
• Add more code generation templates
• Enhance UI/UX with additional themes
• Write comprehensive tests
• Improve documentation
• Report bugs and suggest features

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Ollama for providing the local LLM runtime
• VS Code for the excellent extension API
• The open-source community for inspiration and support

📞 Support

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Email: Create an issue for support

🗺️ Roadmap

Completed in v1.1 ✅

• [x] AI Code Completions (ghost text)
• [x] Quick Edit with Cmd+K
• [x] Multi-file editing support
• [x] Code Review mode
• [x] Undo/Redo system for AI edits
• [x] Prompt Templates library
• [x] Code Snippets library
• [x] Theme toggle (light/dark)
• [x] Thread favorites
• [x] Message reactions
• [x] Performance metrics display
• [x] Advanced context gathering (@terminal, @diagnostics, etc.)
• [x] History search

Coming Soon

• [ ] Voice input support
• [ ] Model comparison view
• [ ] Custom user templates
• [ ] Extension marketplace for agent tools
• [ ] Collaborative coding sessions
• [ ] Integration with popular dev tools
• [ ] Mobile companion app
• [ ] Advanced diff view with multiple algorithms

⭐ Star History

If you find this extension useful, please consider giving it a star on GitHub!

Made with ❤️ by Nishant Unavane

Powered by Ollama • Privacy-First • Open Source