Frequently Asked Questions
Overview
Recure is a free, AI-powered dataset discovery platform. Instead of bouncing between Kaggle, HuggingFace, and other sources, you search once and find what you need, enriched with standardized metadata, quality scores, and side-by-side comparison.
Yes. We are currently in beta. The platform is live and actively growing, and we’re actively adding new sources and features based on community feedback.
Data Sources
Our AI agent pulls metadata from Kaggle and HuggingFace, with Papers With Code and GitHub coming soon. We aim to cover every major open-data platform.
No. We do not download dataset files, only metadata. Every listing links directly to the original source.
How It Works
It runs weekly, it scans connected sources for new and updated datasets, pulls the raw metadata, and enriches it, generating standardized summaries and tags, classifying each dataset by task and modality, normalizing fields like license, size, format, and last updated date, and assigning a health score based on five quality dimensions.
You describe what you need in natural language. Results are matched by intent and ranked by relevance and quality. You can also filter by domain, modality, task, license, size, and more.
Quality & Evaluation
Every dataset on Recure gets a health score. A score from 0–100 across five dimensions: metadata completeness, usability, popularity, freshness, and legal safety.
Tags like "ready-to-use," "well-documented," "commercial-use," or "actively-maintained" so you can assess quality at a glance without opening every listing.
Comparison
Compare 2 datasets across classes, modalities, tasks, license, file types, health score, and source links in one unified view.
Community & Feedback
You can suggest datasets, report issues, or request features. Use the button below to submit feedback directly.
Yes, dataset owners can request updates to their listings by submitting a feedback request with the relevant details.
Dataset Requests
Describe what you need and we'll search across sources and partner communities. We'll keep you updated on progress.
Roadmap
We are adding more sources with deeper analytics, visual previews, benchmark tracking, richer comparisons, and community-driven quality signals.
Subscribe to our Substack newsletter and follow our changelog for the latest product updates and feature announcements.