// specification directory
Dataset specifications
Choose a modality to see the sourcing requirements, rights checks, and deliverables. These are services, not inventory listings.
// sourcing scope
Modality requirements
Language is more than a tag. Define dialect, domain, and the conditions your model needs to handle.
| Modality | Language scope | What to specify | Documentation |
|---|---|---|---|
| Speech & audio | Hebrew · Arabic · dialects | Sample rate, speakers, alignment | Consent & training permissions |
| Video | Spoken language & captions | Frame rate, coverage, timestamps | Source & caption rights |
| Domain text | Hebrew · Arabic · multilingual | Domain, tokens, structure, refresh | Provenance & usage scope |
Illustrative language previewsנתוני דיבור / heبيانات الكلام / arDialect and script verified during evaluation.
Compile your requirements
Compile a readable specification and download it as JSON, or send it for a sourcing review.