Voice Recording ( US Accent, General American)

Role Overview

Mercor is collecting audio accents for a multi-lingual AI platform. Mercor is collecting ~3 hours of phone audio where speakers spell out names, emails, and phone numbers naturally — like talking to customer support. There will be different variations, environments you will need to work with (quietly, muffled voice, Phone on speaker-mode etc.)

How It Works

• You’ll call a demo line and record short sessions (caller side only).
• Each call should sound natural — vary tone, pace, and include small pauses or filler words (“umm,” “let’s see,” etc.).
• Must be real phone calls (no VoIP). Include some background noise for realism.

Requirements

• Must be based in USA, Canada or UK (the phone numbers you will call will be for these geos, Mercor will not reimburse for international calling)
• Must have a General American US accent

Deliverables

• Upto 30 phone calls in 6 variations (~180 calls)
• The code provided by the agent during the call
• Enter call data on a sheet shared with you (meta data, phone call code etc.)
  • Speaker ID or name:
  • Gender:
  • Age group:
  • Primary language:
  • Geographic ethnicity:

What Mercor Needs From You

• Your turnaround time will be 3 hours of conversation that needs to be filled before 12/28 (28th December)