One is the Got It AI platform, which has a system called AutoRAG that includes a TruthChecker model to spot hallucinations and correct them. It's designed for enterprise use and comes with features like graph and metadata generation, automated assessment and security compliance. It's designed to help customer service and sales teams get useful and trustworthy results.
Another is Deepchecks, which automates LLM assessment to spot hallucinations, wrong answers and bias. It combines a "Golden Set" approach with manual overrides to create a richer ground truth for LLM use. It's for developers and teams that want to ensure their LLM-based software is high quality and reliable.
For a full-stack option, check out LastMile AI. It's got tools like Auto-Eval to automatically spot hallucinations, RAG Debugger to optimize performance and AIConfig to optimize prompts and model parameters. It supports multiple AI models and has a notebook-like environment for prototyping so you can more easily generate production-ready generative AI applications.