What the work involves

You record your own voice against structured briefs. A typical batch mixes read speech — sentences, wordlists, digit strings, place names, brand names — with prompted spontaneous speech, where you talk for a set duration on an assigned topic without a script. Some batches ask for specific conditions: a particular speaking rate, a whispered or raised register, a phone-call-style distance from the mic, or a second take with deliberate background noise. Each submission is checked against the brief before it is accepted.

The practical bar is consistency rather than performance. Reviewers listen for a clean noise floor, no clipping, no room echo, stable mic distance across takes, and speech that actually matches the prompt text including hesitations and self-corrections when those are requested. You will also be responsible for basic file hygiene: correct sample rate and format, one utterance per file where specified, accurate naming, and completed metadata about your accent region, age band, and recording setup.

What the platform screens for

  • Verifiable UK English speech. Screens generally include a live or recorded speaking sample. Expect questions about where you grew up, where you have lived since, and how you would describe your own accent region — vague answers tend to fail.
  • A workable recording environment. A quiet room you control, a decent USB or XLR microphone, and headphones. Laptop built-in mics and shared or noisy spaces are usually disqualifying.
  • Instruction fidelity. Whether you follow a brief exactly, flag ambiguity instead of guessing, and can explain why a take should be rejected.
  • Throughput realism. How many usable minutes you can deliver per week, and whether you can sustain that across a multi-week batch.

Logistics

Fully remote and asynchronous. Batches open with a deadline rather than a schedule, so you choose your recording hours — most contributors work in short sessions to protect vocal consistency. Volume is project-driven and uneven: expect bursts of available work rather than steady weekly hours. Pay is typically per accepted deliverable, which means the $10–30/hr observed range depends heavily on how often your takes pass QC on the first pass. Rejected or re-requested audio is generally unpaid.