Diktator! Lite
- Free in this release
- Right Command hotkey
- Recommended local model
- Automatic language detection
- Useful Lite version, not a time-limited trial
Future upgrade
Pro will add power-user control.
Lite is intended to remain a useful free version. Pro is a future one-time purchase plan for people who want more control after the Pro features are real.
Diktator! Pro
- Model and language selection
- Hotkey selection
- Automatic translation to grammatically correct English text
- Local history
- New local models, languages, dictionaries, and writing styles
- Long-form dictation for lengthy texts
- Future optional audio capture workflows with explicit user consent
One-time purchase.
Lite is intended to remain useful without a subscription. Pro is planned as a future one-time purchase when its features are real. Lite does not record the screen or system audio.
In testing
Advanced dictation layer.
The next advanced version is being shaped around natural dictation, not raw transcription: adaptive pauses, context between fragments, live draft text while you speak, spoken formatting, and a final cleanup pass for more readable text. This dictation flow is planned to stay local on the Mac; cloud transcription is not part of it.
Roadmap
Pro v1: Control and Memory
Pro v1: Control and Memory
Installed model picker, additional local model manager, language selection, safe hotkey picker, automatic translation of spoken dictation into grammatically correct English text, searchable local history, and a user dictionary for names, places, projects, technical terms, and multilingual vocabulary.
Pro v2: Output workflows
Plain, cleaned, email, formal, message, and notes styles, with raw and final text stored separately when history is enabled.
Pro v3: Long dictation
Longer dictation sessions for bigger documents, with progress states, recovery, and local handling designed for sustained writing.
Pro v4: Audio Capture archive
Future transcription from macOS Audio Capture for conversations on your Mac would be optional, user-started, and consent-based, with local saving of audio and text only when you choose it.