$50–100/hr · Mercor · Hourly
A native German voice artist records voice samples for AI text-to-speech voice cloning in a customer service agent.
What you would do
- Record voice samples across diverse scripts covering conversational, narrative, instructional, and varied emotional tones
- Deliver multiple takes of the same content with variation in emotion, emphasis, style, and pacing
- Maintain consistent vocal performance across extended recording sessions and potential follow-up sessions
- Ensure technical audio quality by managing microphone setup, room acoustics, and file formatting
- Follow detailed guidance on pronunciation, intonation, and delivery style for each script segment
Who they want
- Native German speaker currently based in Germany
- Clear, natural, and expressive voice with strong control over tone and pacing
- Access to professional or near-professional recording setup with quality microphone and pop filter
- Strong command of German intonation, diction, and emotional expression
- Can commit to one ~4-hour recording session, with potential for a follow-up session of similar length
Main skills
What the interview asks about
1.Consistent vocal delivery
Voice cloning requires stability; if your tone or accent drifts between takes, the model learns an unstable pattern that sounds unnatural.
For example: “Record 'Guten Tag, wie kann ich Ihnen helfen?' in Take 1 conversationally with warm tone, in Take 5 formally. After dozens of other takes, is your accent and voice quality identical in both versions?”
2.Emotional range control
The customer service agent needs to respond appropriately to different customer moods; the TTS model learns to shift emotions by studying your varied takes.
For example: “You are given a customer service response: 'Ihr Problem wird sofort gelöst.' Record it five ways: empathetic and warm, professional and neutral, urgent and energetic, patient and calm, and slightly apologetic. Describe how your voice changes in each version.”
3.Technical audio proficiency
Incorrectly set up microphone distance, poor pop filter use, or background noise ruins training data; the client cannot use flawed recordings.
For example: “Your recording space has ambient traffic noise outside. You have a pop filter, a condenser microphone, and headphones. Walk through the exact setup changes you would make to maximize audio quality for the recording session.”
4.Script fidelity with natural delivery
The model learns to follow text while sounding human; if you ad-lib or rush through scripts, the cloned voice becomes unreliable.
For example: “You are given a script with specific punctuation and a 2-second pause marked at a specific point. Explain how you deliver that pause and what makes it different from simply pausing whenever you feel like it.”
A task you may get
You record three scripted segments with variation requests. You deliver each segment in two different emotional registers, maintaining consistent German accent and microphone technique. Audio quality, consistency between takes, and range are evaluated.
How to prepare
- Test your microphone setup in your recording space and confirm audio levels, background noise, and file format compatibility
- Practice recording the same German script multiple ways to demonstrate emotional and stylistic range
- Review the types of dialogue used in customer service scenarios to prepare for conversational naturalness
- Ensure your pop filter and microphone positioning achieve professional audio quality without harsh consonants or plosives
The facts
- Pay
- $50–100/hr
- Hours
- Hourly
- Where
- Remote
- Open to
- DEU, DEU
- Field
- Language and Audio
- Posted
- 6/1/2026
- Places left
- 9
We wrote this page from the public Mercor listing. It may be out of date, so read the full posting before you apply.