Skip to main content
​
Feature

​
AI-Ready Data is available in beta

Generate evaluation datasets from your own source documents. Upload up to 10 PDF, Word, Markdown, or JSON files, describe what they contain and what should be generated from them, and track the generation as a job.A generated file appears in Seekr, ready to add as a dataset. This feature is in beta. Review a generated dataset before you use it in an examination.See Generate AI-ready data.
​
Launch

​
SeekrGuard is available

Evaluate the safety, quality, and risk of models against data you provide, scored with criteria you define.Register models from OpenAI or Seekr, upload datasets from a file, HuggingFace, or Seekr, and score outputs with the evaluator library or your own LLM-as-a-Judge evaluators. Run examinations across all or part of a dataset, compare a single response across judge models, or chat with a model and score its answers inline. Weight risk categories into a risk profile to turn those scores into a risk rating per model.See What is SeekrGuard.
Last modified on September 8, 2026