Azerbaijani audio to text: what to expect
Azerbaijani sits in the middle range for the speech model. A clear recording of one or two people gives a usable draft, but plan to read it through carefully. Proper names, numbers and fast colloquial speech are where most corrections happen, and noise or music lowers the quality further.
Upload the recording as it is: MP3, M4A, WAV, OGG, OPUS, FLAC, a voice message or a phone recording. You can also record from your microphone in the upload box. Nothing needs converting first.
The transcript opens next to a player that highlights the current line. Click a line to hear it again, correct the word in place, and your changes save automatically.
Alphabet and the closeness to Turkish
The transcript is normally written in the Latin alphabet used in the Republic of Azerbaijan, with ə, ğ, ı, ö, ş, ç, ü, q and x. There is no option to output the Cyrillic script used before the 1990s or the Perso-Arabic script used by Azerbaijanis in Iran.
Azerbaijani and Turkish are close, and the model has much more Turkish data. Auto-detect can mistake a short Azerbaijani clip for Turkish, and even with Azerbaijani selected a few words may come out with Turkish spelling, for example e where Azerbaijani writes ə. Keep Azerbaijani selected and correct those as you listen.
Russian, Persian and regional speech
In Baku many speakers mix Azerbaijani with Russian, sometimes inside a sentence. Russian words may be written in Latin letters as they sound, or misheard as Azerbaijani words. Speakers from Iran (South Azerbaijani) use more Persian vocabulary and a different accent; the model can transcribe them, but expect more corrections than for Baku speech.
- Mostly Azerbaijani with some Russian: keep Azerbaijani selected.
- A long interview section entirely in Russian: fix it by hand or transcribe that part separately with Russian selected.
- Mostly Turkish with Azerbaijani guests: choose Turkish for that file.
Common uses and exports
Typical files are journalism and research interviews, university lectures, podcasts, mugham and poetry evenings, business calls and voice messages from relatives. Students transcribe thesis interviews so they can quote them, and content creators prepare Azerbaijani subtitles. Diaspora families in Russia, Turkey and Georgia send voice notes that people want to keep as text.
Export TXT, DOCX, PDF, CSV, SRT or VTT, with timestamps and speaker names switched on or off per export. Speaker labels are available on paid plans or with a Minute Pack.
For video files, including YouTube interviews, see the Azerbaijani video to text page.