Academic, market and user experience research
Interview transcripts that arrive without the speakers' personal data.
Rota transcribes the interview audio and replaces names, CPF numbers, phone numbers, emails and addresses with placeholders. The material is ready for analysis and for the team, with less manual anonymization work.
Why anonymize at the transcription step
Research ethics committees ask for protection and anonymization of participants' data. Doing this by hand, one interview at a time, takes time and lets details slip through.
When a transcript goes to another person or to an AI tool, the participants' data goes with it, unless someone removes it first.
Placeholders keep people distinct: the same name gets the same placeholder in the whole text, and the analysis of themes and statements remains possible.
How it works
- Send the audio Send the recording, up to 25 MB per file, to the transcription route. Long interviews can be split into parts.
- Receive the clean text The transcript comes back with placeholders instead of names, documents, contact details and addresses. The audio file is not stored.
- Analyze with or without AI Use the text in your analysis tool, or ask a Rota model to group themes and summarize statements.
- Review before publication Read the result and check indirect clues, such as profession, age and city, which Rota does not remove.
Example with fictitious data
Before Rota
Entrevistadora: Pode se apresentar? Participante: Sou a Cláudia Mendes, tenho 42 anos e sou enfermeira há 15 anos. Se precisar, meu telefone é (85) 90000-9753 e meu e-mail é claudia.mendes@example.com.
After Rota
Entrevistadora: Pode se apresentar? Participante: Sou a [PESSOA_1], tenho 42 anos e sou enfermeira há 15 anos. Se precisar, meu telefone é [TELEFONE_1] e meu e-mail é [EMAIL_1].
Fictitious text in Portuguese, processed by the API engine on 6 October 2026, with the placeholder policy.
How to integrate
In the dashboard, you can test with no code: it cleans an excerpt of up to 8 KB and transcribes an audio file of up to 8 MB. To transcribe the interviews every day, call the transcription route with a command such as this one:
POST /v1/audio/transcriptions
curl https://rota-nacional.ia.br/v1/audio/transcriptions \
-H "Authorization: Bearer $ROTA_API_KEY" \
-F file=@entrevista-01.mp3 \
-F language=pt \
-F response_format=textGood to know
- For recordings above 25 MB, send the audio in parts.
- Do a final reading of the material before publication.
FAQ
Frequently asked questions
Direct answers about this use. Did not find yours? Talk to the team.
Is the transcript already anonymized?
It comes without the data that identifies the person directly: names, CPF and CNPJ numbers, emails, phone numbers and addresses become placeholders. The final reading before publication stays with the team.
Is the audio stored?
No. The audio file is not stored. When auditing is active, the input and the output stay encrypted for the defined period, with restricted access.
Do I need to code?
To test it, no: the dashboard cleans an excerpt of up to 8 KB and transcribes an audio file of up to 8 MB, right in the browser. To process the interviews every day, you must run an API command, such as the one in the example above.