v0.2.0

• added support for streaming responses with --stream flag
• new --model option to switch between different llm providers
• fixed token counting issue with multi-byte unicode characters
• refactored prompt handling to support conversation history
• improved error messages when api rate limits are hit

v0.1.0

• initial release of llm-cli with support for openai and anthropic models
• add streaming responses for real time output
• implement conversation history tracking with sqlite backend
• support for custom system prompts via config file
• fix token counting accuracy for gpt-4 models
• refactor provider interface for easier extension