What the work involves

You receive German audio — read speech, spontaneous conversation, interviews, sometimes noisy or multi-speaker recordings — and produce a transcript that is faithful to what was actually said, not to what standard written German would have said. Beyond the raw text, most tasks ask for structured layers: speaker labels, timestamps at defined intervals or turn boundaries, tags for inaudible segments, crosstalk, non-speech events, and occasionally dialect or code-switching markers. A second stream of work is review: taking someone else's transcript and correcting it against the audio, which is judged on whether you catch real errors rather than restyle acceptable choices.

Every project runs on a written style guide, and the guide is the arbiter. It will tell you whether to render hab's verbatim or normalize it, whether to write numerals or words, how to handle Anglicisms, filled pauses, false starts, and Swiss or Austrian variants. Most disagreements in QA come from transcribers applying instinct where the guide has already decided. Where the guide is genuinely silent, the expectation is that you flag it and ask rather than invent a private convention and apply it inconsistently across fifty files.

What the screen looks for

micro1's process is AI-led and moves quickly. Expect questions in and about German that test whether your ear is genuinely native-level: dialect recognition (Bairisch, Schwäbisch, Kölsch, Swiss German, Austrian regionalisms), handling of dropped syllables and regional lexis, and what you do when a speaker is only partly intelligible. You should be able to describe transcription tools you have actually used, your typing speed and realistic audio-hour-to-work-hour ratio, and give concrete examples of guideline calls you have made. Vague claims of "fluent German" and "attention to detail" without a worked example are the most common reason candidates stall.

Logistics

  • Remote, contractor, asynchronous — you pull batches and return them by a deadline rather than working fixed shifts.
  • Observed pay for this band is $10–25/hr, varying with task complexity and review versus first-pass work; rates are not guaranteed.
  • Volume fluctuates with client demand; treat it as supplementary rather than a stable full-time load.
  • You will need a quiet working environment, reliable internet, good headphones, and willingness to sign confidentiality terms covering client audio.