Models
Small, specialized models that run on the device, one per task.
Available
Word timestampsAlign
Accurate word timestamps for any transcript.
Speech enhancementClear
Studio sound, no cloud bill.
Clip selectionClips
Create video shorts and highlights.
Spoken language detectionEar
Detect language based on 30 seconds of audio.
Emoji suggestionsEmo
Suggest emoji faster than you can type.
Image and video understandingEye
Tag and rank images based on aesthetics and content.
Content topic taggingGist
Generate topics and tags for posts and articles.
Content moderationModerator
Flag nudity before upload or showing.
PII redactionRedact
Filter PII on-device.
Shape recognitionShapes
Rough sketch. Perfect shape.
Titles and descriptionsTitle
Suggest a title and description for any text.
Language identificationTongue
Detect language based on 3 words.
Filler-word detectionUhm
Detect and remove filler words in seconds.
Speech recognitionVoz
Transcribe 10 minutes of audio in 2s on an iPhone.
Who said whatWho
Label who said what in audio and video.
