We are building a remote team of native or near-native Swahili speakers to correct AI-generated transcripts. This is careful, detail-oriented language work. The best candidates are people who love the language, notice when something sounds off, and take pride in getting it right.
You will receive audio files alongside a machine-generated transcript. Your job is to listen to the audio and correct the transcript: fixing errors, standardising formatting, and applying our style guide before the file is approved for delivery.
Volume: Flexible, you choose how many files you take on per week.
Turnaround: Files have individual deadlines (typically 2–3 working days from assignment).
Must have
✓ Native or near-native Swahili speaker with strong written skills, bilingual in English
✓ Prior transcription, subtitling, or translation experience
✓ Comfortable with at least one Swahili dialect or register (Nairobi, standard Kenyan, or Tanzanian)
✓ Reliable internet connection and a computer or laptop
✓ Able to follow a detailed style guide consistently
✓ Available for at least 60 audio minutes of work per week
Nice to have
✓ Familiarity with Nairobi Sheng
✓ Exposure to multiple Swahili dialects
✓ Experience with similar transcription platforms
✓ Background in linguistics, journalism, or media
We pay per audio minute of completed and approved transcript. Payment is processed weekly for all files approved in that week.
Typical earnings: An editor working 2–3 hours per day on light-edit files typically completes 60–90 audio minutes per day. Exact earnings depend on file mix and your working pace.
This is freelance work. You are not an employee. We do not guarantee a minimum number of files per week. Volume varies by project. Some weeks will be busy; others slower. We aim to give you reasonable notice of upcoming batches.