v0.0.4

Release Notes

New Features

• AI Provider Selection Enhancements: Introduced a fallback mechanism to enhance the AI provider selection process. The system now defaults to Anthropic when available, with an optional flag to override. Additionally, the --provider flag is now available, supporting options such as OpenAI, Anthropic, and Ollama.
• Anthropic API Support: Added support for the Anthropic API with integration of the Claude model.
• OpenAI Model Selection: Users can now specify custom OpenAI models through CLI flags.

Refactoring

• Code Optimization:
  • Replaced the deprecated io/ioutil package with the io package to modernize the codebase.
  • Simplified provider and model selection logic for clearer flags and defaults.
  • Extracted the OpenAI client and shared Message type into separate files for better modularity.

v0.0.3

Release Notes

Features

• Added support for Ollama as an alternative AI backend, with dynamic client selection based on environment variables
• Enhanced error handling with AI-suggested fixes in JSON format
• Added color formatting to command output and prompt for better readability
• Enhanced command description detail in AI prompts for more informative explanations
• Added service and model announcement on startup
• Added prompt for including command output in conversation context for better context awareness

Improvements

• Enhanced command generation to use structured JSON response format
• Improved JSON parsing error handling
• Updated prompt system to enforce strict response formatting
• Improved Ollama API response handling with content accumulation
• Reorganized imports and removed unused code

Bug Fixes

• Resolved type mismatch errors in command response handling
• Prevented unintended exit after successful command execution
• Fixed handling of Ollama streaming response parsing
• Updated OllamaResponse struct to properly handle created_at as string
• Corrected the addition of assistant's responses to message history

Technical Updates

• Enhanced error logging for Ollama API interactions
• Improved response handling and parsing mechanisms
• Updated command generation workflow for better reliability

v0.0.2

Version 0.0.2

v0.0.1

initial release

• queries openai for commands
• allows followup
• attempts to fix commands based on the error response