Arabic interview transcription for journalists and researchers

Record with a microphone close to the guest, then upload the file to Vocla, choose their dialect and add the names of people and places. You get a text split between the two speakers with timestamps; what remains is checking names and figures and matching quotes against the audio before you use them.

The transcript is a good starting point, not a substitute for checking. On clear recordings most words are right, but any quote published under someone's name should be one you have listened to yourself. That is why every line is tied to its time: click, listen, and correct the word if needed.

Transcripts are kept in your workspace and are reachable by the people you gave a role in it. The protections in place, and what is not done yet, are set out plainly on the security and trust page; read it against your organisation's or ethics board's requirements before uploading sensitive interviews. We do not list certifications Vocla does not hold.

Yes, you choose the dialect on upload and the speech is written in it. Accuracy on dialects is lower than on Modern Standard Arabic and drops further with noise or phone audio, so plan a longer review for field interviews recorded in open places.

Export the interview as CSV, with one row per segment including speaker and time, then open it in Excel or import it into the analysis software you use and add a column for codes. Or export Word if your team codes directly on the text. The coding and the analysis are your work.

Yes, the two speakers are detected automatically and each line is attributed to one of them. It works best when they take turns; interruptions and overlapping speech can end up on a single speaker, which you fix by splitting the line and moving it in the editor.