ThinkReview - Version 1.4.0 Release Notes
Release Date: November 14, 2025
---
🎯 What's New
Local Ollama Integration 🖥️
Complete Privacy & Offline Code Reviews: ThinkReview now supports local AI code reviews using Ollama! Your code never leaves your machine, giving you complete privacy and the ability to review code offline.
- 🔒 Complete Privacy: Your code never leaves your machine - perfect for sensitive projects
- 💰 No API Costs: Free local inference with no usage limits
- 🌐 Offline Capability: Works without internet connection once Ollama is set up
- 🎯 Multiple Models: Choose from specialized code review models optimized for different use cases
- 🔧 Easy Setup: One-click script to start Ollama with proper CORS configuration
Why This Matters:
For developers working with sensitive code, proprietary projects, or in environments with strict data privacy requirements, local AI reviews provide a secure alternative to cloud-based solutions. You get the same intelligent code analysis without sending your code to external services.
Supported Models:
- codellama (4GB) - Fast and efficient for general use
- codellama:13b (8GB) - Better quality for comprehensive reviews
- deepseek-coder:6.7b (4GB) - Excellent code understanding
- qwen2.5-coder:7b (5GB) - Strong code analysis
- starcoder2:15b (9GB) - Great multi-language support
- codegemma:7b (5GB) - Fast and capable
---
⚡ Performance Improvements
Seamless Provider Switching
- Instant Switching: Toggle between Cloud AI (frontier models) and Local Ollama with a single click
- Smart Fallback: Clear error messages guide you if Ollama isn't running
- Connection Testing: Built-in connection test ensures Ollama is properly configured before use
Technical Improvements:
- New OllamaService class handles all local AI interactions
- Automatic CORS configuration detection and guidance
- Model discovery and refresh functionality
- Optimized prompt engineering for local models
---
🔧 Additional Improvements
- Setup Script: New start-ollama-with-cors.sh script automatically configures Ollama for browser extension use
- Comprehensive Documentation: Complete setup guide (OLLAMA_SETUP.md) with troubleshooting tips
- Settings UI: Enhanced popup interface with Ollama configuration section
- Model Selection: Dropdown to select from installed Ollama models
- Connection Status: Real-time connection testing and status indicators
- Error Handling: Improved error messages with actionable suggestions
---
📝 Usage Tips
1. Quick Start with Ollama:
- Install Ollama from ollama.ai/download
- Pull a model: ollama pull codellama
- Run the setup script: ./scripts/start-ollama-with-cors.sh
- Configure in extension settings: Select "🖥️ Local Ollama" and choose your model
2. Choosing the Right Model:
- Start with codellama for fast, general-purpose reviews
- Use qwen2.5-coder:7b for more comprehensive analysis
- Try starcoder2:15b for multi-language projects
3. Hybrid Approach:
- Use Ollama for routine reviews and sensitive code
- Switch to Cloud AI (frontier models) for complex analysis or when offline isn't needed
4. Troubleshooting:
- If you get a 403 error, make sure Ollama is started with CORS enabled
- Use the provided script or set OLLAMA_ORIGINS="chrome-extension://*" environment variable
- Check the setup guide for detailed troubleshooting steps
---
🐛 Bug Fixes
- Improved error handling for Ollama connection failures
- Better validation of Ollama configuration settings
- Fixed model refresh functionality in settings
---
⚠️ Breaking Changes
None - This is a fully backward-compatible release. Existing Cloud AI functionality remains unchanged.
---
📞 Support
Having issues or suggestions? We'd love to hear from you:
- Bug Reports: Click the "Report a 🐞" button in the review panel
- Feedback Form: Share Your Feedback
- Ollama Setup Help: See OLLAMA_SETUP.md for detailed instructions
- Rate Us: Leave a review on the Chrome Web Store
---
Thank you for using ThinkReview! 🚀
Empowering developers with privacy-first AI code reviews.