About our speech and audio data collection
We record real people in real conditions, to the exact mix of languages, accents, ages and environments your speech model needs. Every recording is checked automatically for loudness, noise, clipping and silence before a reviewer listens to it.
Where it's used
- Speech recognition (ASR)
- Voice assistants and IVR
- Text-to-speech voices
- Speaker and accent coverage
What we offer
Read and scripted speech
Prompts and sentences read aloud, for speech recognition and text-to-speech.
Conversational speech
Natural two-person and group conversations, including call-centre style dialogue.
Wake words and commands
Short commands and wake phrases recorded across devices, distances and rooms.
Noisy and real-world audio
Speech in cars, streets, kitchens and offices, with the conditions recorded as metadata.
From first call to final delivery
Scope
We agree the data, the quality bar, the formats and a sample size, and write it into clear guidelines.
Pilot
A small batch first. You review it, we tune the guidelines, and only then do we scale.
Produce
Vetted contributors work in our studios. Automatic checks catch problems as files arrive.
Verify
Consensus between annotators, hidden check items and reviewer audits on every batch.
Deliver
Clean exports in your format, with documentation and a quality report.
What you receive
- Files in your format (e.g. WAV/FLAC, JPG/PNG, MP4)
- Per-file metadata: speaker or contributor profile, device, conditions
- Consent records for every contributor
- Quality report: automatic checks and reviewer decisions