Captions
Speech from your Mac's system audio becomes stable captions in a small, movable bubble. Recognition runs on the Mac.
sAiity for Mac
Speech on the machine, not in the cloud.
Live captions, translation, transcripts, and press-to-talk dictation for macOS.
macOS 26 or later Apple Silicon Windows not supported
- Processing
- On-device
- Account
- Not required
- Platform
- macOS / Apple Silicon
Four ways to keep speech useful.
Translation
Add a second line in another language with a local translation model while keeping the original speech available.
Dictation
Hold a configurable key, speak, and release. Silence-aware regions preserve language switches, while local cleanup removes clear fillers and repetitions. The result goes into the focused field when possible and always remains on the clipboard.
Transcripts
Save original and translated sessions in the app, then read them later or export SRT, WebVTT, Markdown, or plain text.
Local by default
Nothing leaves your Mac for daily use.
There is no account, tracking, telemetry, or audio upload. Models run on-device after download. The captions microphone is optional; dictation opens it only while the press-to-talk key is held.
Current release
Ready when you are.
GitHub Releases is the canonical download location. Sparkle checks the signed feed only when you enable updates. Version 2.6.0 adds an optional local translation engine, safer dictation shortcuts, readable subtitle exports, faster dictation cleanup, and fixes for model lifecycle, caption timing, and transcript search.