Skip to content
trainsetpartners Build a brief
Menu

// Dataset brief compiler

Compile a sourcing brief

Tell us the training data you need. Include modality, language, rights, and scale.

Dataset brief builderREQUEST / V1.0
Dataset requirements

Use hours for audio/video, documents or tokens for text, and items for images.

Illustrative Prefilled values are an example. Edit to match your model.
Rights and delivery constraints
Contact and review

Review your requirements before sending.

Send requirements only. Do not include datasets, sensitive records, credentials, or confidential contracts.

What happens after you submit

  1. Our sourcing team reviews the brief within one business day.
  2. We match your requirements to qualified datasets we have rights to use.
  3. Within two business days we send a summary of potential sources, any open questions, and a proposed timeline.
  4. If you decide to proceed, we provide a detailed quote and a confidentiality agreement before any data exchange.

What to include in your brief

  • Modality – speech, audio, video, or text.
  • Language & dialect – e.g., Modern Hebrew, Gulf Arabic, English (US).
  • Rights needed – training only, retrieval, both.
  • Scale – approximate number of hours, minutes, or documents.
  • Quality expectations – transcription accuracy, speaker diversity, background noise level.
  • Refresh frequency – one‑time delivery or ongoing updates.
  • Domain focus – medical, legal, finance, technical, etc.

What not to include

  • Full data files or large attachments.
  • Proprietary source code or models.
  • Confidential contracts that are not yet signed.
  • Personal identifiers beyond what is required for rights verification.

Keep the first step small

Provide estimates, not exact counts. Use “Not sure” for any rights question you cannot answer. Distinguish what you know from what still needs checking.

Need more context?

Read how we work at /delivery/. Explore our specialty services: /specifications/speech-and-audio-data/, /specifications/video-training-data/, /specifications/domain-specific-text-data/. Browse recent case studies at /guides/.