Features Overview
WYN360 CLI provides comprehensive AI-powered development assistance through a rich set of integrated features.
Core AI Capabilities
🤖 Interactive AI Assistant
- Natural Language Interface - Describe what you want in plain English
- Conversation Memory - Maintains context across multiple interactions
- Multi-Provider Support - Anthropic Claude, Google Gemini, AWS Bedrock
- Model Switching - Change AI models mid-session based on task complexity
📝 Code Generation & Analysis
- Production-Ready Code - Generate complete applications, not just snippets
- Smart File Operations - Context-aware file creation, reading, and updates
- Intent Recognition - Understands "update existing" vs "create new" automatically
- Self-Correcting - Smart retry mechanism with error handling
Agentic Features (v0.4.0)
🧠 Memory & Skills
- Persistent Memory - Remember context across sessions (user preferences, project decisions, feedback)
- Relevance Selection - Automatically surfaces the most relevant memories per query
- Custom Skills - Define reusable slash commands via YAML files in
~/.wyn360/skills/ - Skill Aliases - Multiple names for the same skill (e.g.,
/review,/cr)
📋 Planning & Sub-Agents
- Plan Mode - Structured step-by-step plans with approve/reject/skip workflow
- Parallel Workers - Spawn sub-agents for concurrent research and verification
- Task Lifecycle - Track worker status (pending, running, completed, failed, killed)
- Result Synthesis - AI combines parallel findings into coherent next steps
🔧 Hooks & Token Budget
- Hook System - Pre/post hooks for validation, transformation, and custom processing
- Safety Hooks - Built-in detection of destructive commands (rm -rf, DROP TABLE)
- Auto-Continue - Automatically continues when responses are cut off by token limits
- Diminishing Returns - Detects when continuations produce little new content and stops
Advanced Features
🌐 Web Integration
- Real-Time Web Search - Access current information and resources
- Direct Website Fetching - Read specific URLs directly
- Autonomous Browsing - AI navigates websites and extracts data
- Smart Caching - Efficient content caching with TTL management
📄 Document Processing
- Multi-Format Support - Excel, Word, PDF, and more
- Vision Mode - Process images, charts, and diagrams
- Semantic Chunking - Intelligent document segmentation
- Embedding Search - Find relevant content across large documents
🔧 Development Tools
- GitHub Integration - Commit, push, create PRs seamlessly
- HuggingFace Deployment - Deploy to Spaces with one command
- Automatic Test Generation - Generate comprehensive pytest tests
- Command Execution - Run any CLI tool with safety confirmations
Feature Comparison by Provider
| Feature | Anthropic Claude | Google Gemini | AWS Bedrock |
|---|---|---|---|
| Code Generation | ✅ Excellent | ✅ Very Good | ✅ Excellent |
| Web Search | ✅ Built-in | ⚠️ Custom tool | ✅ Built-in |
| Document Processing | ✅ Full support | ✅ Full support | ✅ Full support |
| Vision Mode | ✅ Advanced | ✅ Good | ✅ Advanced |
| Browser Automation | ✅ Supported | ✅ Supported | ✅ Supported |
| Context Window | 200K tokens | 2M tokens | 200K tokens |
| Cost per M tokens | $3.00/$15.00 | $0.075/$0.30 | $3.00/$15.00 |
| Best for | Complex reasoning | Cost efficiency | Enterprise use |
Development Workflow Features
📊 Session Management
- Conversation History - Review past interactions
- Token Tracking - Monitor API usage and costs in real-time
- Session Save/Load - Preserve important conversations
- Model Information - View current model, pricing, and capabilities
⚙️ Configuration & Customization
- Hierarchical Config - Project, user, and default settings
- Custom Instructions - Add personal coding standards to every conversation
- Project Context - Help AI understand your tech stack
- Environment Variables - Runtime configuration overrides
🎯 User Experience
- Multi-line Input - Shift+Enter for newlines, Enter to submit
- Streaming Responses - See results as they're generated
- Rich Terminal UI - Beautiful formatting with syntax highlighting
- Slash Commands - Quick access to features and settings
Security & Privacy
🔒 Credential Management
- AES-256 Encryption - Secure storage of API keys and tokens
- Per-User Keys - Isolated encryption for multi-user systems
- Audit Logging - Track authentication attempts without exposing secrets
- Session Management - Secure cookie storage for web authentication
🛡️ Safety Features
- Command Confirmation - User approval required for system commands
- Timeout Protection - Prevents infinite loops and runaway processes
- File Permissions - Respects system security boundaries
- Error Isolation - Failed operations don't crash the session
Performance & Optimization
⚡ Speed Features
- Streaming Responses - Start reading while AI is generating
- Smart Caching - Website and document content caching
- Parallel Processing - Multiple operations can run concurrently
- Token Optimization - Intelligent context management
💰 Cost Management
- Multiple Models - Choose based on task complexity and budget
- Token Limits - Configurable limits to control costs
- Usage Tracking - Real-time monitoring of API consumption
- Cost Transparency - Separate tracking for different API types
Integration Ecosystem
🔗 Platform Integrations
- GitHub - Repository management, PRs, issues
- HuggingFace - Space deployment, model hosting
- AWS - Bedrock integration for enterprise users
- Docker - Container management and deployment
🛠️ Developer Tools
- Poetry - Python dependency management
- Pytest - Automated testing framework
- Playwright - Browser automation
- Git - Version control operations
Feature Roadmap
Current Focus (v0.5.x)
- ✅ Dream (auto memory consolidation)
- ✅ Context compaction
- ✅ Vim mode
- ✅ Voice input (speech-to-text)
- ✅ Buddy companion system
- ✅ Cron/scheduled agents
- ✅ Plugin system
- ✅ LSP diagnostics integration
- ✅ Conversation rewind
Previous (v0.4.x)
- ✅ Persistent memory system
- ✅ Sub-agent parallel workers
- ✅ Structured plan mode
- ✅ Token budget auto-continue
- ✅ Extensible skills system
- ✅ Hook-based pipeline
Previous (v0.3.x)
- ✅ Multi-provider AI support
- ✅ Autonomous browsing capabilities
- ✅ Document processing with vision
- ✅ Google Gemini integration
Upcoming Features
- Claude Computer Use - Direct computer interaction capabilities
- Enhanced Vision - Advanced image and diagram processing
- Team Collaboration - Shared sessions and project contexts
Get Started with Specific Features
- Dream & Compaction → - Auto memory consolidation and context management (v0.5.0)
- Vim, Voice & Buddy → - Vim editing, speech input, virtual companion (v0.5.0)
- Cron, Plugins & LSP → - Scheduled agents, plugins, diagnostics, rewind (v0.5.0)
- Memory & Skills → - Persistent memory and custom slash commands (v0.4.0)
- Planning & Sub-Agents → - Structured plans and parallel workers (v0.4.0)
- Hooks & Token Budget → - Pipeline hooks and auto-continue (v0.4.0)
- Web Search → - Find current information and resources
- Browser Use → - Autonomous web navigation and data extraction
- Vision Mode → - Process images and visual content
- GitHub Integration → - Seamless repository management
- HuggingFace → - Deploy and share your projects
Next: Explore specific features or check out Usage Examples →