framework-caching

What problem does it solve?

Caches for PDF processing pipelines to improve performance by storing semantic results, embeddings, and LLM outputs.

Core Features & Use Cases

• Semantic caching for search results to avoid recomputing similarities on repeated queries.
• Embedding caching for request vectors to accelerate repeated analyses.
• LLM caching to reuse model outputs for identical prompts within TTL windows.
• Invalidation and TTL controls via environment variables for predictable freshness.
• CLI and API integration for monitoring and management.

Quick Start

Enable the caches by setting CACHE__SEMANTIC_ENABLED, CACHE__EMBEDDING_ENABLED, and CACHE__LLM_ENABLED in your environment, then run your PDF workflow to see caching in action.