What the work involves
This is a single-session audio contribution task, not an ongoing engagement. You receive a short English script and a step-by-step recording guide, record yourself reading it aloud in your own voice, and upload the file through the platform. Start to finish, contributors report about 15 minutes including reading the guide, doing a test clip, and checking the recording before submission.
The purpose is to build speech data representing authentic Danish-accented English — the kind of accent variation that speech models routinely mishandle. That makes authenticity the whole point. The recording must reflect how you actually speak English as a Danish speaker, not a performed or exaggerated accent, and not a neutralised one either.
What the platform screens for
- Genuine linguistic background. Screens probe where you learned Danish and English, how long you have lived in Denmark or a Danish-speaking environment, and whether Danish is a first or heritage language. Imitated accents are explicitly excluded.
- Instruction-following without hand-holding. The guide covers file format, ambient noise, mic distance and take structure. Submissions that ignore a step are typically rejected rather than returned for revision, so the screen looks for evidence you read written specs carefully.
- Usable recording conditions. A smartphone is fine. A noisy shared space, a fan running, or a room with heavy echo is not.
Logistics
Fully remote and asynchronous, with no scheduled calls or fixed hours. Turnaround is short — these projects tend to open and close within days, so responsiveness after acceptance matters more than availability across a week. Pay is a flat $50 per accepted submission rather than hourly; there is no guarantee of follow-on work, though contributors on similar AfterQuery language batches are sometimes invited back when a new locale batch opens.