Validate LLM responses from various models with AI supervision.

Looking for a single endpoint to build testing libraries or manual scoring of responses with larger models. Replacing human review equivalent for simpler response tasks and categorizations.

Basic API for internal use

Does not necessarily need a GUI can just be an end point with various fetches to OpenAI for example or using other testing frameworks. Should provide option to select best completion, grade 1-10, process batch analysis and summarize. Ideally looking at what the very best libraries and techniques are.

Open to best simple approach

Open to best simple technical approach. Python or Node.js