The AI Playground supports two types of playgrounds:
1
The File Playground is used to analyze a single document at a time. It is ideal for use cases where information needs to be extracted, validated, or summarized from one file, such as an invoice, contract, ID card, or report. Users can upload a document, test different prompts, and evaluate how effectively the AI responds to queries on that individual file.
2
The Dataset Playground is designed for working with a collection of documents grouped together as a dataset. Instead of analyzing files individually, the AI evaluates information across all documents within the dataset at once. This is useful for scenarios where information must be compared, validated, or analyzed across multiple files, such as verifying whether several KYC documents belong to the same person, comparing contracts, or identifying inconsistencies across records. It helps users test prompts that rely on relationships and context between multiple documents rather than a single file.