Axion AI
VS Code Extension | Your intelligent AI coding assistant
Your intelligent AI coding assistant for Visual Studio Code
50+ AI Models • 13 Specialized Modes • Free AI Options • Zero Compromise Privacy
GitHub Repository
Features •
Installation •
AI Models •
Modes •
Usage
Features
Multi-Provider AI Support
Connect to 50+ AI models from leading providers:
Premium Providers:
- Anthropic - Claude Opus 4, Sonnet 4, 3.7/3.5 Sonnet, Haiku
- OpenAI - GPT-4.1, GPT-4o, O3 Pro, O3, O1 (reasoning models)
- Google - Gemini 2.5 Pro/Flash, Gemini 2.0, 1.5 Pro/Flash
- xAI - Grok 3, Grok 2, Grok Vision
- DeepSeek - V3, R1 (reasoning), Coder
- Mistral - Large, Medium, Small, Codestral, Pixtral
Free AI Providers (Get Free API Keys):
- Groq - Llama 3.1 70B/8B, Mixtral 8x7B, Gemma 7B (Free tier - Ultra fast inference)
- Hugging Face - Llama 3, Mistral 7B, CodeLlama, Phi-2 (Free token - Open source models)
- Together AI - Llama 3 70B, Mixtral 8x7B, Qwen 2.5 (Free tier - Multiple models)
- Perplexity - Sonar, Sonar Online (Free tier - Research-focused)
Beautiful Modern UI
- Sleek dark theme with glass morphism effects
- Real-time streaming responses
- Syntax-highlighted code blocks
- Smooth animations and transitions
- Responsive design that fits your workflow
Codebase Awareness
- Reads and understands your entire project
- Context-aware suggestions based on your code
- Smart file tree analysis
- Active file content awareness
Smart Code Editing
- Propose file changes with one click apply
- Undo/revert changes easily
- Copy code blocks instantly
- Insert code at cursor position
Conversation Management
- Save and load chat history
- Multiple conversation support
- Delete old conversations
- Persistent storage across sessions
Secure & Private
- API keys stored locally in VS Code settings
- No data sent to third parties
- Direct connection to AI providers
- Your code stays private
Installation
From Source (Development)
- Clone or download this repository
- Open the folder in VS Code
- Run
npm install to install dependencies
- Run
npm run compile to build
- Press
F5 to launch Extension Development Host
From VSIX
- Download the
.vsix file
- In VS Code: Extensions →
... → Install from VSIX
- Select the downloaded file
- Reload VS Code
Modes
Axion AI offers 13 specialized modes for different tasks:
Choose the right mode for your workflow - from simple Q&A to advanced multi-AI collaboration
Agent Mode
Full coding capabilities with read/write access
- Analyze and understand your codebase
- Propose and implement code changes
- Create new files and refactor existing code
- Fix bugs and add features
- All changes require your approval
Ask Mode
Read-only assistant for learning and exploration
- Explain code and concepts
- Answer questions about your project
- Provide insights and suggestions
- No file modifications allowed
Library Coder Mode
Specialized for library development
- Design reusable APIs
- Create well-documented code
- Follow library best practices
- Generate examples and documentation
- Support for any programming language
Advanced Modes
- Swarm - Multi-AI swarm intelligence consensus
- Evolution - Code evolution through generations
- Quantum - Quantum code optimization
- Neural - Neural architecture search
- Time Travel - Time travel debugging
- Predictive - Predictive coding assistant
- Mind Meld - Deep style understanding
- Parallel - Parallel universe testing
- DNA - Code DNA analysis
- Collaborative - Multi-AI collaboration
Usage
Getting Started
Open Axion AI - Click the Axion AI icon in the activity bar or press Ctrl+Shift+K (Mac: Cmd+Shift+K)
Configure API Keys - Click Settings and enter your API keys:
- Free providers: Get free API keys from Groq, Hugging Face, Together AI, or Perplexity
- Premium providers: Add your API keys for Claude, GPT, Gemini, etc.
Select Model - Choose your preferred AI model from the dropdown
Pick a Mode - Select from Agent, Ask, Library, or any of the advanced modes based on your task
Start Chatting - Type your question or request and press Enter
Keyboard Shortcuts
| Shortcut |
Action |
Ctrl+Shift+K / Cmd+Shift+K |
Open Axion AI |
Ctrl+Shift+N / Cmd+Shift+N |
New Chat |
Enter |
Send message |
Shift+Enter |
New line in message |
Getting API Keys
Premium Providers (API Key Required):
| Provider | Get API Key |
|----------|-------------|
| Anthropic | console.anthropic.com |
| OpenAI | platform.openai.com |
| Google | aistudio.google.com |
| xAI | x.ai |
| DeepSeek | platform.deepseek.com |
| Mistral | console.mistral.ai |
Free Providers (Optional API Key for Better Limits):
| Provider | Get Free API Key | Notes |
|----------|------------------|-------|
| Groq | console.groq.com | Free tier available - Ultra fast inference |
| Hugging Face | huggingface.co/settings/tokens | Free token - Open source models |
| Together AI | together.ai | Free tier available - Multiple models |
| Perplexity | perplexity.ai | Free tier available - Research assistant |
Note: Free providers offer free API keys with generous rate limits. Get your free API key from the provider's website to start using these models immediately.
Applying Code Changes
When in Agent mode, Axion AI proposes changes in a special format:
- Review the suggested code in the chat
- Click Apply to apply the change
- Click Revert to undo if needed
- Changes are applied directly to your files
Configuration
Configure via VS Code Settings or the extension settings panel:
{
"axion-ai.anthropicApiKey": "sk-ant-...",
"axion-ai.openaiApiKey": "sk-...",
"axion-ai.googleApiKey": "AIza...",
"axion-ai.xaiApiKey": "xai-...",
"axion-ai.deepseekApiKey": "sk-...",
"axion-ai.mistralApiKey": "...",
"axion-ai.groqApiKey": "gsk_...",
"axion-ai.huggingfaceApiKey": "hf_...",
"axion-ai.togetherApiKey": "...",
"axion-ai.perplexityApiKey": "pplx-..."
}
Tip: Start with free providers! Get free API keys from Groq, Hugging Face, Together AI, or Perplexity to use Axion AI immediately without any cost.
Supported AI Models
Access the latest models from top AI providers - all in one place
Anthropic (Claude)
| Model |
Description |
| Claude Opus 4 |
Most capable, complex reasoning |
| Claude Sonnet 4 |
Best balance of speed & quality |
| Claude 3.7 Sonnet |
Extended thinking capability |
| Claude 3.5 Sonnet/Haiku |
Fast and efficient |
| Claude 3 Opus/Sonnet/Haiku |
Previous generation |
OpenAI (GPT & O-Series)
| Model |
Description |
| GPT-4.1 / Mini / Nano |
Latest GPT-4 models |
| GPT-4o / Mini |
Multimodal (vision + text) |
| O3 Pro / O3 / O3 Mini |
Advanced reasoning |
| O1 Pro / O1 / O1 Mini |
Reasoning models |
Google (Gemini)
| Model |
Description |
| Gemini 2.5 Pro/Flash |
Most capable, thinking mode |
| Gemini 2.0 Flash/Lite |
Fast multimodal |
| Gemini 1.5 Pro/Flash |
Long context |
xAI (Grok)
| Model |
Description |
| Grok 3 / Fast / Mini |
Latest Grok models |
| Grok 2 / Vision |
Previous gen with vision |
DeepSeek
| Model |
Description |
| DeepSeek V3 |
Latest chat model |
| DeepSeek R1 |
Reasoning model |
| DeepSeek Coder |
Code specialist |
Mistral
| Model |
Description |
| Mistral Large/Medium/Small |
General purpose |
| Codestral |
Code specialist |
| Pixtral Large |
Vision support |
Privacy & Security
Your code never leaves your machine. API keys are stored locally. No middlemen.
- Local Storage: All API keys are stored locally in VS Code settings
- Direct Connections: Code is sent directly to the AI provider you select
- No Middleman: No data passes through any intermediate servers
- Your Control: Conversations stay on your machine
Building for Distribution
# Install vsce globally
npm install -g @vscode/vsce
# Package the extension
vsce package
# This creates axion-ai-1.0.0.vsix
Contributing
Contributions are welcome! Feel free to:
- Report bugs
- Suggest features
- Submit pull requests
License
MIT License - Use and modify freely!