Citrusx

Citrusx validates machine learning and LLM models, explains their decisions and alerts teams the moment a model makes mistakes in production. It is aimed at regulated companies where data science and model risk teams must show models are fair and compliant.

Data scientists use the web UI and SDK to check model performance and investigate results, while managers compare models and track KPIs. Risk officers get reports to assess model soundness against regulation, and executives receive summarized custom reports.

The platform is model-agnostic and also covers retrieval-augmented generation (RAG) pipelines in LLM applications. Access is arranged through a demo with the sales team.