Some of these roles ask you to agree that your voice may be cloned. The details below come from Mercor’s data-policy page and the open audio postings on Mercor and micro1, read 13 October 2026.
What the cloning notice says, and what it leaves out
Of the 24 open audio roles, 13 are Mercor voice-cloning sessions. Every one carries a flagged line saying your voice may be cloned for the client’s customer-service AI agent. The line asks you to apply only if you accept that. Every one also says the recordings are used only for that agent and are not sold, licensed or reused elsewhere.
That is a stated limit on use. The postings still leave gaps:
- How long the recordings and the cloned voice are kept. None of the 13 postings says.
- Whether you can withdraw consent later or ask for deletion.
- Who owns the recordings, and whether you are paid again if the client uses the voice for longer.
- What happens if the client changes, or the lab behind it does.
Mercor’s data-and-AI-usage page covers your account data, including interview audio and transcript. It says deleting your account removes them, and also that deletion does not remove data from companies that already accessed it. It says nothing about voice clones. The micro1 German and English recording posting tells you to follow the project’s consent and data guidelines, which it does not print.
Ask for the consent text in writing before you record. This site cannot advise on voice, likeness or privacy law where you live. If you have an agent or a union membership, ask whether either covers synthetic copies of your voice. None of these postings mentions either.
Recording sessions are booked by accent and by country
The 13 voice-cloning postings cover these accents and locales: UK English, Swiss German, Australian English, African American, Southern USA, Standard French, German, Northern UK, Spanish (Peninsular), Spanish (Mexico), Spanish (Latin America), Canadian French and USA. Of these, 12 ask you to be based in the country or region concerned. The US alone is required by 5 of the 24 audio roles.
All of them describe one session of about four hours, with a second of similar length possible but not guaranteed. Mercor calls them rolling openings. The UK English posting wants a neutral, standard accent. The Southern USA and Northern UK postings name example regions instead, such as Texas and Georgia, or Yorkshire and Liverpool. The Northern UK posting asks for female voice actors. Prior voice acting is a plus and not required.
Three other kinds of audio work
Reading and acting at volume. micro1’s German and English recording posting asks you to read sentences in both languages and record controlled background noise from a stationary spot. You also transcribe clips and tag the files. The Spanish (Mexico) voice actor posting is heavier. It asks for 20 or more hours a week, a demo reel and an acoustically treated room. The interview happens on your real recording setup.
Listening. Mercor’s Audiobook QA postings, 8 of them, ask native speakers who listen to audiobooks to flag problems in AI narration. The list includes skipped words, misread numbers, wrongly expanded abbreviations and odd intonation. They run about 20 hours a week and can be based anywhere. micro1’s Voice Actor/Coaches posting asks you to describe accent, resonance, pitch and room sound in written reports, and to bring good headphones.
Music transcription. micro1’s Audio Expert posting is not about speech. It wants SATB choral lines separated from a mono mix and written as MusicXML, with Python. It states an hourly range but says pay is output-based, per accepted task, with a weekly minimum number of tasks.
What the roles pay
| Kind of work | Roles | Advertised band, an hour |
|---|---|---|
| Voice-cloning recording sessions | 13 | $50–150 |
| Rating voices (micro1) | 1 | $30–70 |
| Audiobook review | 8 | $15–25 |
| Music transcription (micro1) | 1 | $50–100 |
| Standing language and audio pool (Mercor) | 1 | $35–50 |
Recording sessions state $50–150 an hour, against $15–25 for audiobook review. Within audiobook review, the English (US) project states $20–25 and the other 7 state $15–20. A recording session is about four hours at the stated hourly rate. The postings give no session total, so this site does not multiply one out.
The kit the postings expect
Mercor’s postings ask for a professional or near-professional setup: a quality microphone, a quiet room and a pop filter. They name Audacity, Adobe Audition and Reaper as tools worth knowing. The Mexico posting names an audio interface, a microphone and headphones as preferred kit. No posting states a sample rate, bit depth or file format. They refer to detailed recording guidelines that the postings do not print.
The listening roles ask for a careful ear and, in micro1’s case, good headphones.
What the postings do not say
- How many takes or scripts a session holds, and how long you wait for pay after it.
- Why an application is turned down, and how long a rolling opening stays open.
- Whether the second session ever comes. Mercor calls it possible, not promised.
- How the standing language and audio pool turns into paid work. Mercor says contact can come within a week or several months.
This site is not affiliated with any platform named here, and passing a screening is not a promise of work.
Open voice and audio roles right now
The 6 newest of 24 open roles this site lists today, with the pay each posting states.