Show HN: CacheCore – semantic LLM caching with dependency invalidation (Rust)

Category: infrastructure

Tags: llm-caching, semantic-cache, rust

Score: 7.3/10 (Innovation: 7, Technical: 7, Documentation: 5, Utility: 8)

CacheCore is a semantic caching proxy for LLM agents that detects redundant reasoning steps across multi-agent pipelines, reducing API costs by 40–70%. It integrates as a drop-in OpenAI-compatible endpoint, making it interesting for teams running production agent workloads with frameworks like LangChain or CrewAI.

Target audience: backend devs, ai engineers, devops

Repository: https://cachecore.it/

View on Hacker News