For AI companies
Specialised business data, rights verified and documented.
We aggregate data from many small and mid-sized businesses into datasets you can license under one agreement, with provenance for every source.
Data categories
What we source.
- Product catalogs
- Structured product descriptions, attributes and specifications across sectors.
- Technical documents
- Manuals, procedures, standards and specification sheets written by domain specialists.
- Images
- Original product and process photography with business-held rights.
- Video
- Demonstrations, training material and operational footage.
- Audio
- Owned recordings, available with transcripts where they exist.
- Written content
- Long-form articles, reports and editorial text authored in-house.
We also take requests for categories not listed here. Describe what you need below.
Multilingual coverage
Data in the languages your models need.
Our suppliers are businesses writing for their own customers, in their own languages. That makes their material a source of natural, domain-specific text in languages that are thin in public web data.
Tell us the languages and domains you need. We source to the request, and the dataset card states the language breakdown of what is delivered.
Documentation
Delivered with every dataset.
Rights verification is the core of what we do. Each delivery includes the records your legal and data governance teams will ask for.
Provenance record
For every source: which business contributed it, how the material was created, and the basis on which they hold the rights. Supplier identities can be disclosed or kept confidential depending on the agreement.
License terms
A single license covering the full dataset, with permitted uses, restrictions, term and territory stated plainly.
Anonymisation method
What categories of personal data were found, how each was removed or transformed, and what was excluded.
Dataset card
Composition, languages, volumes, formats, known gaps and collection period.
Formats
JSONL, CSV or Parquet for text and structured data; raw files with a manifest for media. Other formats on request.
Delivery
Secure transfer to your storage, with checksums and a manifest of every file.
Process
Request, sample review, agreement, full delivery. Sample sets can be provided before you commit.
Dataset request
Describe what you need.
We reply by email with what we can source and how we'd document it.