Show HN: Nexa-gauge – Cache/cost-aware graph-based eval for LLM and RAG

Category: devtools

Tags: llm-evaluation, rag, cache-aware, graph-based, cli-tool, python

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

Nexa-gauge is a graph-based evaluation engine for LLM and RAG systems that provides cache-aware execution, cost estimation, and structured reports. It stands out by combining deterministic metrics with LLM-as-a-judge in a reusable, repeatable pipeline that minimizes API spend. This is particularly interesting for teams needing rigorous, cost-controlled evaluation in production workflows.

Target audience: backend devs, data engineers, ml engineers

Repository: https://github.com/harnexa/nexa-gauge · Python · MIT · 8 stars

View on Hacker News