📝Auto Captions

Select "Auto Captions" to instantly generate animated captions.

Characters per caption

Choose how many characters appear in each caption.

Unique settings are saved for each Sequence aspect ratio: Landscape, Portrait, and Square.

Model

Choose transcription models with varying speed & accuracy.

Parakeet

NVIDIA Parakeet provides superior transcription for European languages.

Supported languages: English, Bulgarian, Croatian, Czech, Danish, Dutch, Estonian, Finnish, French, German, Greek, Hungarian, Italian, Latvian, Lithuanian, Maltese, Polish, Portuguese, Romanian, Russian, Slovak, Slovenian, Spanish, Swedish, Ukrainian

Name	Size	Description
Parakeet v3 Compact	396MB	Fast transcription for European languages.
Parakeet v3	637MB	High-quality transcription for European languages.
Parakeet v3 HQ	1.17GB	Max accuracy for European languages.

Whisper

Whisper provides transcription for all major languages.

See all 99+ supported languages

Name	Size	Description
Tiny	41.5MB	Fastest model with basic accuracy.
Base	78MB	Balanced model offering good accuracy and reasonable speed.
Small	252.2MB	High accuracy model with moderate speed.
Medium	785.2MB	Very high accuracy model with slower processing but excellent results.
Large V3 Turbo	833.7MB	Most accurate text transcription with slower processing time. Recommended for non-English languages.
Large V3	2.9GB	Highest accuracy model. Best for offline/batch transcription where speed is not critical.

Transcription models are saved locally to your machine. Delete and redownload models at any time.

Windows

C:\Users\Public\Captioneer\models

MacOS

~/Users/*username*/Library/Application Support/Captioneer/models

Language

Select Auto to auto-detect the spoken language or choose a transcription language.

Selecting a transcription language speeds up processing time.

Single-word captions

Check this box to create fast-paced transcription that appears one word at a time.

Also generate a Caption Track

Check this box to create a new Caption Track after processing.

VAD (Voice Activity Detection)

Choose from 4 presets for precise speech detection with Whisper models: Auto, Podcast, Monologue, Fast-paced. Choose Custom to dial in your own advanced controls.

Processing Details

Captioneer lists the available GPU and models on your machine. Choose between Automatic, GPU, or CPU processing.