What the work involves

You'll record voice samples from supplied scripts — conversational turns, narrative passages, instructional copy, and the kind of support-line phrasing a CX agent needs to handle. Sessions are self-directed: you receive scripts and recording specs, record in your own space, check the takes, and upload files in the required format. Expect to do multiple passes on the same lines with varied emotion, emphasis, and pacing, since TTS training benefits from controlled variation rather than a single perfect read. Consistency across weeks matters more than any one heroic session — the model is learning your voice, so drift in accent, mic distance, or room tone becomes a data problem.

What the platform screens for

Mercor's screen is AI-led and will focus on things it can verify: that you are Australia-based with a native Australian accent, that you have real voice acting, dubbing, narration, podcasting or broadcast history you can point to, and that your recording setup is genuinely quiet and clean. Expect to be asked about your microphone, interface, treatment, and how you handle plosives and background noise. You'll almost certainly be asked to submit or perform a read, and to demonstrate range across at least a couple of registers — calm corporate, warm conversational, brisk instructional. The consent question is explicit and non-negotiable: voice cloning is the point of the project, and candidates who are unsure about that should not apply.

Logistics

  • Fully remote, recorded from your own space, on your own schedule within deadlines
  • Roughly 5–10 hours per week for the project duration; reliability across the whole run is weighted heavily
  • Must be currently based in Australia
  • Pay observed at $50/hr for this listing; Mercor rates vary by project and are not guaranteed
  • Familiarity with Audacity, Audition, or Reaper is preferred, mostly for checking and trimming your own takes rather than heavy post-production