What the work involves
You receive audio or video clips in Portuguese — read speech, spontaneous conversation, call-centre recordings, accented or noisy field audio — and produce a verbatim or lightly cleaned transcript against a written style guide. Guides differ by client: some want filled pauses, false starts and stutters retained; others want them stripped. You will also spend time on the review side, correcting machine-generated transcripts (ASR output) rather than typing from scratch, which is a different skill — spotting a plausible-sounding hallucinated word is harder than hearing a blank.
Typical task types:
- Verbatim transcription with speaker diarisation and timestamps
- Post-editing ASR output for word error rate reduction
- Tagging non-speech events, crosstalk, and inaudible segments per convention
- Dialect and register labelling (pt-BR vs pt-PT, regional variants, formality)
- Adjudicating disagreements between two prior transcribers
What the platform screens for
micro1's screening is AI-led and conversational. Expect it to confirm your native or near-native Portuguese, ask which variety you actually work in and how confidently you handle the other, and then push on specifics: orthography under the Acordo Ortográfico, how you would render a mispronounced proper noun, when a segment is genuinely inaudible versus when you are guessing. It also checks the unglamorous logistics — headphones, quiet room, stable connection, whether you can hit a turnaround window. Vague answers about "fluency" get followed up until they resolve into something checkable.
Logistics
Fully remote and asynchronous. Work arrives in batches with deadlines rather than fixed shifts, so hours are flexible but uneven — a busy week may offer 20+ hours, a quiet one none. Pay is hourly or per audio minute depending on the project; the $10–18 band reflects rates contributors have reported, not a guarantee. A closed-back headset and a room where you can listen without ambient noise are practical necessities, not preferences.