v0.4.0

Release notes:

• Updated log formatting
• Enhanced agent base class with more advanced features from a prototype autonomous agent class, including automatic planning, output validation, recovery, and approvals
• Improved the planner
• Added new examples and fixed old ones
• Implemented an overridden parallel run in the advanced agent team to work with the advanced team features
• Merged agent router classes into one class as well as the 'advanced' agent team into one class, removing the 'advanced' separate class files

v0.3.0

Release notes:

• Improved planner implementation
• Improved agent plan execution, especially with tool usage
• Added advanced agent team sequential run method
• Added function_tool interop between parentheses usage vs none
• Implemented autonomous agent (WIP - may or may not make it to 1.0 release)
• Added new examples

v0.2.2

Documentation update

• Updated README.md with correct package installation instructions (name change)

v0.2.1

Initial PyPI release as agentix-lib

Changes:

• Package renamed to agentix-lib for PyPI compatibility
• Set up GitHub Actions for automated publishing
• Configured PyPI trusted publisher integration

v0.2.0 - Fixes, improvements, and new features & tools

Release notes:

• Multiple fixes to a few implementations, more notably with the base Agent class and how it handles providing final answers properly
• Several improvements including a refactor of LLM convergence checking, tool handling overhaul, and additional failsafes
• Several new features and tools including a metrics logger, Tavily tool, DuckDuckGo tool, Firecrawl tool, YFinance tool, function calling tool decorator, new LLM provider integrations, and more examples

v0.1.0 - Initial Release

Release notes:

• Initial release of the Agentix framework
• Provides the core foundation and advanced configurations for implementing LLM agents
• Basic and advanced usage examples and documentation are now available