Viewing profile — sulci_cache

Building Sulci — context-aware semantic cache for LLM apps.

Blends conversation history into cache lookups so "What are the main differences?" resolves correctly based on prior turns. +20.8pp accuracy over stateless caching. 0.74ms hit latency.

Patent pending. Apache 2.0.

https://sulci.io https://github.com/sulci-io/sulci-oss pip install "sulci[sqlite]"