Just talk. Help tech hear every voice.
Pair up for easy two-person conversations — some from a short script, some free-flowing — indoors and out around town. We supply the body camera and every script and prompt. A few sessions and you're done, and voice technology works better for people of every background, age and accent.
- 200 spots
- open to participants
- ~30 min
- of your speech each
- 50 / 50
- indoor and outdoor
How a session works.
Every recording is a conversation between two people, captured on a body camera we hand you.
- 01
You're paired up
Every recording is a conversation between two participants — never a solo recording. You can be paired with anyone; varied, natural pairings are preferred.
- 02
Pick up your batch
Each group works through one batch of materials. We provide everything: scripted dialogs to read, and prompts to talk about. You don't write anything.
- 03
Record, indoors and outdoors
A chest-mounted body camera, supplied by us, records the conversation as one continuous file. Roughly half your sessions are indoors and half outdoors — streets, parking lots, parks, plazas and sidewalks.
- 04
Repeat until you hit ~30 minutes
A 30-minute conversation yields about 15 minutes of speech per person, so you'll take part in several sessions. You're done once your own speech reaches roughly 30 minutes (27–33 is fine).
Two kinds of conversation
Scripted and free-form.
Your time is split roughly evenly between indoors and outdoors, and both settings include both kinds of conversation.
~55%
Two people hold an unscripted conversation in response to a prompt we give you.
~45%
Two people perform a pre-written dialog. Slips, mistakes and stumbles are normal and totally fine — no re-takes needed.
Who can take part.
We're deliberately recruiting a wide range of voices — different ages, backgrounds, accents and regions — because the goal is a model that works well for everyone. The sign-up form asks a few background questions so we can balance the group.
- Ages 16 and up — teens, adults and seniors are all welcome
- A mix of backgrounds, accents and regions across the United States
- Comfortable talking naturally in English with another person
- Able to join several sessions, indoors and outdoors
- Willing to sign a written release covering use of your voice recordings to train and evaluate speech-recognition systems
Recording ground rules
- Use the body camera exactly as supplied — don't change any of its settings
- Talk at a normal conversational distance, standing or seated
- Record each conversation as one continuous take
- Ordinary room noise, wind and traffic are expected — just keep speech understandable
- Speak naturally and keep your own way of talking; never "correct" your grammar or accent
Consent and privacy.
Your voice is personal. Here's what we ask, and why.
Written consent, specific to voice
Every participant signs a release that specifically covers using voice recordings to train and evaluate speech-recognition systems. Generic media releases don't count.
Voice can be biometric data
Some state laws treat voice recordings as biometric identifiers, so consent, retention and deletion are handled with that in mind.
Under 18
Participants aged 16–17 need written consent from a parent or legal guardian in addition to their own agreement. Nobody under 16 can take part.
Final consent paperwork and compensation details are shared with you after you sign up and before any recording begins. Signing up is not a commitment.