# Changelog
All notable changes to this project will be documented in this file.
The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
## [0.9.1] - 2026-01-02
### Changed
- **Code Cleanup**: Comprehensive code review and refactoring
- Removed ~600 lines of dead code (`server_manager.py`)
- Consolidated duplicate `DateTimeEncoder` class into `client.py`
- Consolidated duplicate `__version__` definition
- Made `test_model_responsiveness` method public (was `_test_model_responsiveness`)
### Fixed
- **Configuration**: Default timeout now correctly set to 60s (was 30s in code, 60s in docs)
- **Imports**: Removed unused imports (`json` from hardware_checker, `subprocess`/`time`/`datetime` from advanced_tools)
- **Docstrings**: Fixed inconsistent version references and project name in module headers
### Added
- **Developer Documentation**: New `CLAUDE.md` with project conventions, structure, and patterns
- **Code Comments**: Section markers and algorithm documentation for complex functions
### Technical Details
- All 11 unit tests passing
- No breaking changes to public API
- All MCP tools remain fully functional
---
## [0.9.0] - 2025-08-17
### Fixed
- **Critical Bug**: Fixed datetime serialization issue in model listing
- Resolved `Object of type datetime is not JSON serializable` error
- All datetime objects now properly converted to ISO format strings
- Model listing and all other tools now work correctly with Claude Desktop
### Technical Details
- Fixed datetime handling in `list_local_models` tool
- Improved JSON serialization for all API responses
- Enhanced error handling for edge cases
- All tools tested and verified working with Claude Desktop MCP client
### Testing
- Comprehensive test suite run on Windows 11 with RTX 4090
- All 9 tools verified working correctly:
- ✅ list_local_models
- ✅ local_llm_chat
- ✅ ollama_health_check
- ✅ system_resource_check
- ✅ suggest_models
- ✅ select_chat_model
- ✅ test_model_responsiveness
- ✅ start_ollama_server
- ✅ remove_model
## [Previous Versions]
### [0.8.x] - July-August 2025
- Initial development and architecture setup with Claude AI assistance
- Implementation of core MCP tools
- Cross-platform compatibility framework
- GPU detection and system analysis features
### [Initial Release] - July 2025
- First version by Paolo Dalprato developed with Claude
- Basic Ollama integration and model management