01Five source languages into English
Pick the language spoken in the file — Spanish, German, French, Japanese or Chinese — and English as the target. Telling us the source language (rather than guessing) gives noticeably better transcription on short or noisy files.
02An English transcript you can actually use
Even if you only need the words, the review step is useful on its own: you get the original and the English side by side, with timecodes, and you can correct terms of art before exporting. Sentences with numbers or dates that don't match the source are flagged.
When you approve, the English is voiced and laid on the original timeline. You get both the MP3 and the aligned bilingual text.
03Natural English, not word-for-word
We ask for idiomatic, speakable English that keeps every idea and figure from the source but doesn't mirror its word order. Japanese and Chinese in particular compress meaning into fewer characters, so English versions are often longer; the voice adjusts its pace slightly to stay in sync.
04When it works best
- One main speaker, recorded clearly (phone memos are fine; crowded rooms are hard)
- Spoken content: lessons, talks, podcasts, voice notes, interviews you conducted
- Recordings you own or have written permission to translate
QQuestions
Q.01How accurate is the English translation?
Good enough to publish after a read-through for most clear recordings. That is why the job stops for you to review: you see every sentence before it is voiced.
Q.02Can I just get the English text without audio?
Today a job always ends with audio, but the bilingual transcript is shown on the job page as soon as translation finishes, and you can copy it from there.