Find licensed data for your next AI model.
xchange.ai packages public, open, and creator-licensed datasets into searchable catalog cards with provenance notes, license obligations, AI-readiness grades, and download-ready samples.
Browse → Customize → Select → Train
The enterprise training-data workflow, made self-serve for builders who cannot wait through quote-gated procurement.
Browse
Search dataset cards by domain, modality, locale, license grade, availability, and AI-readiness.
Customize
Start from curated packs for RAG, coding assistants, legal QA, ASR, OCR, geospatial, finance, and biomedical work.
Select
Review provenance summaries, allowed/prohibited uses, attribution files, and sample manifests before you download.
Train
Use signed R2 delivery, checksums, compliance exports, and access logs for your model-development workflow.
Launch catalog preview
First production cards emphasize legal, finance, biomedical, knowledge-graph, speech, code, and public-domain text datasets.
Wikipedia English Articles Dump
General knowledge · Text / encyclopedia · CC-BY-SA 4.0
Project Gutenberg Public-Domain Books
Literature / humanities · Text / books · Public Domain (US)
PubMed Open Access Biomedical Text
Biomedical · Text / XML / abstracts · CC-BY / CC0 / PD mix
arXiv CS / AI / ML Papers
Scientific research · Text / LaTeX / PDF · Mixed open licenses + arXiv terms
CourtListener US Court Opinions
Legal · Text / judicial opinions · Public Domain + Free Law terms
SEC EDGAR 10-K / 10-Q / 8-K Filings
Finance / corporate disclosures · Text / HTML / XML · Public Domain (US government)
Compliance is the product, not an appendix.
Free dataset hubs prove discovery demand. xchange.ai adds the missing buyer layer: curation, license review, attribution manifests, access logs, and audit-ready exports.
No NC/ND production listings
Non-commercial and no-derivatives data stays out of the paid catalog.
License obligations surfaced
Attribution, share-alike, citation, and commercial-use cautions are visible before download.
R2 signed delivery
Large datasets move through zero-egress Cloudflare R2 with time-limited access.
Zero-result learning loop
Search misses become procurement tasks and use-case-pack priorities.
25 use-case packs underway
Published pricing below the enterprise quote wall.
Keep enterprise white-glove quotes, but let indie builders, researchers, and AI teams start without waiting for a sales call.
Browse catalog, license summaries, limited samples, three small public-pack downloads per month.
25 downloads/month, saved collections, attribution files, and low-quota API key for indie builders.
250 downloads/month, larger mirrors, notebooks, license-obligation exports, and priority requests.
Team seats, API metadata access, bulk downloads, signed URLs, compliance exports, private collections.
White-glove search, custom licensing, procurement support, audit package, private delivery options.
Give one to get one.
Contributors who submit a dataset that passes license and provenance review can receive one free access credit for a dataset of equal or lower grade. It is grade-gated so useful data earns useful access — and junk data earns nothing.
Join the submitter queueReady to build with cleaner data?
Start with the catalog, request a pack, or submit supply for review.