Skip to content
aidataset.broker home

Request data

Request an enterprise dataset

Describe the dataset you need: data type, languages, period, rough volume, intended use and licence. We reply by email with what can be assembled and how it would be documented. Datasets are assembled to your spec, de-identified, documented and delivered under a licence with a datasheet. A request costs nothing and commits you to nothing.

A man with glasses on a video call at his desk at night, a notebook of handwritten notes in front of him
Languages

Rough is fine: records, threads, repositories or years of history.

Use
Licence

Sectors, team functions, period, formats, exclusions: anything that defines the spec.

No commitment. We use your details only to answer this request.

FAQ

Request questions

Does sending a request commit me to anything?

No. A request tells us what to look for. Nothing is agreed until you have seen a sample and signed a licence.

What happens after I send a request?

We reply from hello@aidataset.broker and say plainly whether we can assemble the dataset, in what shape, and what we need from you to scope it.

What should the spec include?

Data type, languages and variants, period, sectors or team functions, rough volume after filtering, intended use, licence preference and anything that must stay out. The spec sheets list what matters per type.

Can data vendors and RL-environment builders send requests?

Yes. Requests from AI labs, data vendors and teams building RL environments or evaluations are all welcome. The licence names the uses you agree on.

What will a delivered dataset include?

Files in the published schema, a datasheet, a de-identification report and a rights record, under a written licence.

Keep reading

Spec sheets by data typeWhat to put in your spec.How labs procure enterprise dataSteps from spec to licence.Datasheet templateWhat comes with every delivery.