Licensed datasets,
ready for training.

Curated, versioned, deduplicated datasets for computer vision, NLP and tabular ML — delivered through a single API with clear commercial licensing.

● 1,200+ datasets ● Commercial license included ● S3 / GCS delivery ● EU-hosted
01

Clean & deduplicated

Every dataset passes automated dedup, PII scrubbing and schema validation before publishing.

02

Clear licensing

Commercial-use license bundled with every dataset. No scraping ambiguity, no takedown risk.

03

API-first delivery

Pull directly into your training pipeline via signed URLs, versioned snapshots and webhooks.

Pull a dataset snapshot

Authenticated REST endpoint, returns a signed S3 manifest.

# GET https://api.rom-dataclub.com/v1/datasets/ocr-receipts-eu/latest
curl https://api.rom-dataclub.com/v1/datasets/ocr-receipts-eu/latest \
  -H "Authorization: Bearer $DC_TOKEN"

Popular datasets

Vision

Retail Shelf Detection

142k images, bbox + segmentation

NLP

EU Legal Contracts

38k documents, 6 languages

Tabular

POS Transactions

9M rows, anonymized

Vision

OCR Receipts EU

210k images, 12 countries