Automatic caching for LLM API responses (OpenAI, Gemini, Anthropic) using a lightweight Python library.
Smart text chunker for LLM preprocessing (sections → paragraphs → sentences → hard splits).
