Captions and transcription

Choose a provider#

Saryu offers Apple On-Device, Whisper Local, and Cloud transcription. Apple and Whisper process the audio on your Mac.

Cloud is optional. It uses Deepgram for most languages and Sarvam saaras:v4 for Telugu. Saryu uploads audio through Convex for that request, so Cloud needs an internet connection and a signed-in Saryu account.

Deepgram supplies word timing and speaker labels. Sarvam supplies phrase timing for Telugu, so Saryu estimates each word position inside the phrase. Telugu Cloud captions do not include speaker labels.

Create captions#

Open Captions, choose the language and provider, then create the transcript. Longer recordings take more time.

Review names and specialist words. Speech tools can hear them incorrectly even when the timing is right.

Edit and style#

Select caption text to correct it. Use the caption style controls for font, size, color, placement, and animation.

Keep captions inside the safe frame. Preview the whole video before export.