Resources · Datasets

Datasets you can actually use.

Open + licensed samples curated by Natton — language, vision, audio and agentic data.

Dataset

IndicConverse-22 (sample)

Conversational data sample across 22 Indic languages — open subset.

May 15, 2026
Dataset

Med-DICOM-Anno (sample, de-identified)

De-identified DICOM annotation sample with bounding regions + class labels.

Apr 10, 2026
Dataset

AP-Invoice-1k

1k anonymised invoices for KIE evaluation across SAP + custom layouts.

Mar 18, 2026
Dataset

Enterprise AI Training Dataset

A curated dataset for training and evaluating enterprise AI models across customer support, finance, and operations.

Mar 1, 2026

Ready to build
AI you can trust?

Talk to a solutions architect — get a pilot scoped in 48 hours.