Transcribe Macedonian audio to text

Macedonian is a smaller language with a lot of material still locked in audio: radio programmes, interviews, lectures and local government sessions. A written transcript makes that material searchable and quotable, which is where most people start.

The first thing to know is the Cyrillic alphabet. Macedonian has its own letters, including ѓ and ќ, and they carry meaning, so a transcript that loses them would change words. Because the output is plain editable text, you can proofread these letters and correct anything that looks off.

The second thing is closeness to Bulgarian and Serbian. Speakers sometimes borrow words from neighbouring languages, especially in news and sport. If a name or loanword looks wrong, check it against how the speaker actually said it, using the timestamps to jump to the right moment in the audio.

Macedonian also attaches the definite article to the end of the noun, so you will see forms such as "состанокот" for "the meeting". That is normal and not an error. Accuracy varies with audio quality, accents and background noise, so a short test file is a sensible first step before a long batch.