Greek audio and video to text

Modern Greek, Cypriot and regional speech

The target is modern Greek as spoken in Greece. Cypriot Greek has its own sounds and words, and so do Pontic and Cretan, so expect more mistakes with them. Ancient Greek and katharevousa are not what the model is built for. Mixed Greek and English, common in business and technology, is harder than pure Greek.

Accents, question marks and capitals

Modern Greek uses the single accent (monotonic system), and the transcript follows it, not the old polytonic marks. Check that stress accents are placed on the right syllable, because a wrong one can change the word. The Greek question mark looks like a semicolon (;), so one at the end of a sentence is correct. Greeklish in Latin letters is not produced.

What Greek speakers use it for

University lectures, online classes, meeting notes, interviews, family voice messages, and the Greek diaspora recording relatives' stories. Videos made for a Greek audience can also get captions, which helps people watching without sound. Greek names and place names are worth checking in particular, since they are often written in more than one way, and a transcript you can search makes that easy.

Which model works best for Greek

Greek is supported with a good amount of training material, but less than the largest languages, so the small model makes more slips than it does for English or Spanish, especially on rare words, names and fast or noisy speech. Expect a draft that needs a proper read-through.

On tryscriptorium.com the no-account preview and regular accounts run the small model (the larger medium and large-v3 models run when you self-host Scriptorium). In an account, pick Small in the model menu rather than Auto: Auto prefers the fastest, least accurate model for long files or a busy queue.

The small model handles clear, standard Greek reasonably well. Names, place names and foreign words written in Latin letters inside Greek speech are the usual errors, and stress accents deserve a check. For dialect-heavy audio, use short clips or a larger self-hosted model.

Subtitles and captions in Greek

Greek subtitles typically use two lines of about 35 to 42 characters, and Greek words are fairly long, so give each line enough screen time. Exports are UTF-8, so Greek letters appear properly in YouTube and VLC. In all-capital titles Greek traditionally drops the accents, so keep lowercase for readability.

The uploader is set to Greek

The upload box on this page transcribes as Greek, so the recogniser does not have to guess the language from the first seconds. If the recording is in another language or switches languages, use Audio to text, which detects the language automatically. Automatic transcription is a draft and no accuracy figure is promised: check names, numbers and anything you will rely on.