Free trial for 14 days, no card required. Then Pro Monthly at $19/month, Pro Annual at $179/year (saving $49 vs monthly), or Pro Lifetime at $499 one-time. Educational discount: 50% off Monthly or Annual with a .edu email address. Cancel anytime.
Free beta with 2,000 newest photos, full version $79 one-time purchase
Key features
Hold-to-talk hotkey dictation in Mail, Slack, editors, terminal and browser
Verbatim transcription by default with optional cleanup settings per app
Live partial transcript caption visible while speaking
On-device speech processing on Apple Silicon after model download
Personal dictionary to bias transcription toward names and technical terms
Voice snippets—short trigger words that expand to full text before insertion
Lost-dictation recovery—transcripts and audio preserved if delivery fails
Long-form dictation up to 10 minutes per hold without word limits
Menu-bar native Swift app with no Dock icon or notification popups
Choice of cloud or local transcription with audit log of network activity
Automatic face detection and grouping across library
Map view with camera location data
All processing runs locally on your computer
Hide photos, faces, or people from view
Automatic indexing of new photos
Support for multiple libraries with separate settings
Read-only access to cloud storage
Expert controls for clustering and detection
What makes it different
Verbatim by default—leaves grammar and slang alone, unlike most AI dictation that automatically tidies speech
On-device recognition option on Apple Silicon keeps your speech processing entirely on your Mac
Transparent claims you can verify—signed, notarized, listed frameworks, and audit log visible in the app
Built by one maker as a native Swift app designed to feel natural with the keyboard
Works system-wide in your text fields without floating windows or interruptions
No photos leave your device or are sent to Aubora servers for processing
Desktop app with no server to maintain, unlike self-hosted alternatives
Halopen
Aubora