Transcribe Portuguese audio to text, free

What makes Portuguese hard to transcribe

Brazilian and European Portuguese sound much further apart than they look on paper. European speakers reduce unstressed vowels so heavily that whole syllables nearly vanish, while Brazilian speech keeps them open. A model that handles one well can measurably degrade on the other.

What your Portuguese transcript looks like

Nasal vowels (ã, õ) and the full accent set are written correctly. Spelling follows the variety spoken rather than being normalised.

[00:00 → 00:04] Bom dia a todos, vamos começar a reunião.

[00:04 → 00:08] Primeiro vamos rever os resultados do mês passado.

An illustration of how Portuguese transcripts are formatted — script, punctuation and timestamps. Not the output of any particular recording.

How accurate to expect

Brazilian Portuguese is the stronger of the two varieties given far greater training coverage. European Portuguese, with its heavy vowel reduction, needs cleaner audio for the same result.

What usually needs fixing

1. 1 Vowel reduction in European Portuguese swallowing entire unstressed syllables.
2. 2 Distinguishing the many spellings that map to the same nasal sound.
3. 3 Brazilian regional accents from the Northeast, which are less represented than São Paulo and Rio speech.

Getting better results

• → Brazilian Portuguese is the stronger case of the two, simply because more training audio exists.
• → European Portuguese recorded on a phone at distance is the hardest combination — get closer to the microphone.

Who transcribes Portuguese