OwnKey CLI v1.0.0 - Stable Release

🎉 OwnKey CLI v1.0.0 - Stable Release

We're excited to announce the first stable release of OwnKey CLI! This release brings a production-ready AI coding agent that works seamlessly with or without a database.

🌟 What's New in v1.0.0

Hybrid Storage System

• ✅ Works with OR without database - No Supabase required!
• ✅ Automatic local storage in ~/.ownkey/storage/
• ✅ Seamless fallback if database unavailable
• ✅ 7-day retention with auto-cleanup
• ✅ No data loss - always has a backup

Fully Functional Commands

All commands now work perfectly in both local-only and database modes:

• ✅ ownkey suggest - AI-powered code analysis
• ✅ ownkey apply - Apply suggestions safely with backups
• ✅ ownkey undo - Restore from backups interactively
• ✅ ownkey scan - Fast project scanning
• ✅ ownkey config - Easy configuration

Multi-Provider AI Support

• ✅ Google Gemini (tested & verified)
• ✅ Ollama (tested & verified)
• ✅ OpenAI (tested & verified)
• ✅ Anthropic Claude (tested & verified)

🐛 Bug Fixes

• Fixed UV handle errors causing crashes
• Fixed database disconnection race condition
• Fixed apply/undo requiring database
• Improved error handling across all commands
• Enhanced stability in both local-only and database modes

🚀 Quick Start

# Install
npm install -g ownkey
# Configure (local-only mode)
ownkey config --provider gemini --api-key YOUR_KEY
# Analyze code
ownkey suggest . --local-only
# Apply suggestions
ownkey apply local-123456789-0
# Undo if needed
ownkey undo

v0.8.0 - Ollama Integration

OwnKey now supports local LLMs via Ollama! Run code analysis completely offline without API keys.

New Features

Ollama Support

• 100% Offline - No internet required
• No API Keys - Completely free
• Maximum Privacy - Code never leaves your machine
• Any Model - DeepSeek Coder, CodeLlama, Llama3, and more

Quick Start with Ollama

# Install Ollama
# Download from https://ollama.ai
# Pull a model
ollama pull deepseek-coder

# Configure OwnKey
ownkey config --provider ollama

# Analyze offline!
ownkey suggest .

v0.7.0 - Apply System with Automatic Backups & Rollback

v0.7.0 - Apply System

OwnKey can now automatically fix your code! Apply AI-generated suggestions with automatic backups and instant rollback.

New Features

Apply System

• Auto-Fix Code - Apply AI suggestions to fix bugs, security issues, and more
• Automatic Backups - Every apply creates a timestamped backup
• Interactive Confirmation- Review changes before applying with colored diff preview
• Dry Run Mode - Preview what would change without applying
• Validation - Prevents conflicts and bad applies
• Automatic Rollback - Reverts changes if apply fails

Undo System

• List Recent Applies - See all your recent changes
• Instant Rollback - Undo any applied change
• Selective Undo - Choose specific applies to revert
• Batch Undo - Undo all applies from today

🚀 Usage

Apply a Suggestion

# Interactive mode (recommended)
ownkey apply <suggestion-id>
# Auto mode (no confirmation)
ownkey apply <suggestion-id> --auto
# Preview only
ownkey apply <suggestion-id> --dry-run

v0.6.0 - Multi-Provider Support

Major Features:

• Add OpenAI provider (GPT-4 Turbo, GPT-4, GPT-3.5)
• Add Anthropic provider (Claude Opus 4.5, Sonnet 4.5, Haiku 4.5)
• Implement dynamic model detection via provider APIs
• Support for 15+ AI models across 3 providers
  Technical Changes:
• Extended AIProvider base class for code reuse
• Added async getAvailableModels() to all providers
• Updated AIProviderFactory for multi-provider support
• Fixed 15 TypeScript strict mode errors
  Breaking Changes:
• getAvailableModels() is now async and requires API key parameter
  Defaults Updated:
• Anthropic default: claude-opus-4-5-20251101 (world's best coding model)
• Gemini default: gemini-2.5-flash (higher quota limits)
• OpenAI default: gpt-4-turbo
  Documentation:
• Updated README with all 3 providers
• Added provider comparison table
• Enhanced quick start examples
• Added API key links for each provider

Related: #v0.6.0

"v0.5.0 - Google Gemini AI Integration"

Gemini API integration is working!