Hindi audio and video to text

Hindi mixed with English

Many Hindi speakers mix English words into their sentences. That mixing is a known weak point for speech recognition: the English words may come out spelt in Devanagari or the wrong script, or be misheard. Read those lines against the audio.

Names and spelling

Names and borrowed words have more than one spelling, so the editor is where you make the text match your own convention. Search for a name to find every place it appears.

What it is used for

Interviews, lectures, meetings, family recordings, and video that needs Hindi subtitles. Export the transcript as text or as SRT or VTT.

Which model works best for Hindi

Hindi is among the languages Whisper-family models have seen the least speech in, and the small model is often only a rough guide. Use short, clean, single-speaker clips and correct the draft against the audio. The larger models help when you self-host Scriptorium, but even they make many errors in Hindi.

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.

Hindi is where a larger model would help most, and tryscriptorium.com tops out at small. Output is in Devanagari. If the recording is really Urdu or mixed Hindi and English, expect extra corrections; try a short clip first.

Subtitles and captions in Hindi

Devanagari needs a font that supports it in your video editor. Keep lines to two short lines, and break at a natural pause rather than inside a conjunct word. Subtitles stay in Hindi, the language of the transcript.

The uploader is set to Hindi

The upload box on this page transcribes as Hindi, 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.