Releases · parsaghaffari/browserbee

Release Date: October 22, 2025
Tag: v0.2.0
Previous Release: v0.2.0-beta

🚀 What's New

This major release brings next-generation AI models, comprehensive testing infrastructure, and a completely redesigned user interface to BrowserBee.

✨ Key Features

🤖 Next-Generation AI Models

Claude 4.5 Series: Sonnet 4.5 (new default), Haiku 4.5, and Opus 4.1
- 64K output tokens for Sonnet/Haiku (8x increase!)
- Enhanced extended thinking capabilities
- Improved reasoning and coding performance
GPT-5 Series: GPT-5, GPT-5 Mini, and GPT-5 Nano
- Reasoning model support
- Advanced capabilities for complex tasks
OpenAI-Compatible APIs: Support for custom endpoints and third-party providers

🛠️ Robust Development Infrastructure

100+ Unit Tests: Comprehensive test coverage across all components
CI/CD Pipeline: Automated testing and deployment
Hot Module Reload: Enhanced development experience
ESLint Integration: Code quality and consistency

🎨 Redesigned User Interface

New Options Page: Clean, organized layout with vertical tabs
Getting Started Guide: Built-in onboarding and support
Enhanced Model Management: Better pricing display and configuration
Modular Architecture: Improved maintainability and extensibility

🔧 Enhanced Provider Ecosystem

Ollama Improvements: Custom model management and better local integration
Anthropic Updates: Extended thinking for all Claude 4+ models
OpenAI Enhancements: Reasoning model support and better error handling

📊 Release Statistics

30+ commits since v0.2.0-beta
5 new AI models (Claude 4.5 + GPT-5 series)
100+ unit tests implemented
3 enhanced providers (Anthropic, OpenAI, Ollama)
Complete UI redesign with modular components

🎯 Benefits

For Users

🧠 Smarter AI: Latest models provide more accurate browser automation
🎨 Better Interface: Intuitive, organized options and settings
🔒 More Reliable: Comprehensive testing ensures stable operation
⚡ More Choice: Expanded model and provider options

For Developers

🧪 Test-Driven: Full test suite with coverage reporting
🔧 Modern Tooling: Hot reloading, linting, and CI/CD
📐 Better Architecture: Modular, maintainable codebase
📚 Great Docs: Enhanced guides and contribution instructions

🔄 Migration Guide

From v0.2.0-beta

No breaking changes! This release is fully backward compatible.

Recommended actions:

Update your default model: The new default is Claude Sonnet 4.5 (better performance)
Explore new models: Try Claude Haiku 4.5 for faster responses or GPT-5 for advanced reasoning
Check the new options page: Redesigned interface with better organization

Model Recommendations

General Use: Claude Sonnet 4.5 (new default) - Best balance of speed and capability
Fast Responses: Claude Haiku 4.5 - Near-frontier performance at high speed
Complex Reasoning: Claude Opus 4.1 or GPT-5 - Maximum capability for difficult tasks
Cost-Effective: Claude Haiku 4.5 - Great performance at lower cost

🐛 Bug Fixes

Fixed development server issues and build optimization
Corrected model naming consistency across providers
Resolved pricing display issues for o3 model
Enhanced error handling and response processing

📦 Technical Details

Dependencies Updated

@anthropic-ai/sdk: Updated for Claude 4.5 support
openai: Enhanced for GPT-5 series
Added comprehensive testing stack (Jest, Testing Library)
Updated build tools and development dependencies

Architecture Improvements

Modular component design for better maintainability
Enhanced provider abstraction for easier extensibility
Improved error handling and logging
Better token management and caching

🙏 Acknowledgments

Special thanks to all contributors who made this release possible:

Community feedback and bug reports
Testing and validation efforts
Documentation improvements
Feature suggestions and enhancements

🔗 Links

GitHub Repository: https://github.com/parsaghaffari/browserbee
Chrome Web Store: Available for installation
Discord Community: Join for support and discussions
Documentation: Enhanced guides and API references

Full Changelog: v0.2.0-beta...v0.2.0

Download: Latest Release

New Features

Security & Control Enhancements

Approval System: Added "requires approval" flag to irreversible tools, requiring explicit user confirmation before executing potentially destructive actions

Token Counter: Added token usage tracking to help users monitor their LLM consumption

User Experience Improvements

Better Multi-Window Handling: Improved management of multiple Chrome windows

Fixed "frame detached" error when reloading the extension
Resolved bug where approval dialog appeared in all windows

Context Summarization: Added capability to summarize context for improved agent performance

LLM Integration

Adaptable LLM Support: Added configuration options in the extension settings

Automatic refresh of model list upon configuration changes
Support for displaying the active LLM

Ollama Integration: Added support for local LLM usage through Ollama

Memory Management

Task Memory System: Implemented memory storage using indexedDB

Added memory import/export functions
Included pre-built memories for major websites
Ensured memories are stored/retrieved using canonical URLs

Code Quality

Robustness and Refactoring: Improved code quality and architecture

Documentation: Enhanced project documentation including README, license, and acknowledgements

Marketing Materials: Added promotional videos

Releases: parsaghaffari/browserbee

Release list

BrowserBee 0.2.0 Release Notes