Custom conversational collection

Full-duplex conversational speech data for turn-taking and interruption models

Collect real two-speaker interaction with separate channels, natural overlap, interruptions, backchannels, and timing metadata for speech-to-speech systems and low-latency voice agents.

Free sample: Tell us your target language and dialogue setup. The team selects a matching separate-channel conversation example and sends it manually within 48 hours, free of charge.

Collection specification

Speakers
Paired natural conversation with role and demographic quotas
Channels
Separate speaker tracks, synchronized delivery, optional mixed reference
Interaction
Overlap, interruptions, backchannels, repairs, pauses, and turn transitions
Annotation
Speaker turns, timestamps, overlap intervals, interruption outcomes, and transcripts
Evaluation
Private held-out split available as a separately documented deliverable
Pricing
Custom, scoped to your conditions

Conversation is more than alternating clean turns

Human speakers acknowledge, interrupt, hesitate, restart, and speak at the same time. A model trained only on isolated utterances or neatly alternating dialogue does not learn when to listen, yield, continue, or recover.

Separate synchronized channels preserve those events without forcing diarization to guess who spoke. The annotation scope can stay lightweight or extend into precise turn-taking labels depending on the model job.

This offer is not the right fit when

  • You only need isolated command recordings.
  • Synthetic dialogue is acceptable for every evaluation condition.
  • You need an off-the-shelf fixed corpus with no collection lead time.

Get a free sample

  1. Your details
  2. Your project
  3. Verify

Three short steps. The team then follows up manually within 48 hours, and confirms volume, rights, QA, and delivery if you want the scope priced.

Would rather talk it through first? Price a full-duplex collection

Prepare the brief

Bring a specification your vendors can price consistently

Use the free worksheet to define coverage, rights, delivery fields, held-out rules, and acceptance tests before requesting a collection plan.

Build a dataset specification

Buyer documentation

Specimen data card and provenance structure, evaluation brief template, and acceptance and held-out questions.

Guides for this decision

Need a scoped collection plan? Send the language, channel, volume, annotation, and timing you know. The quote form opens directly in project-brief mode.