No UI. Just say it.
Start by clicking on the status bar button or by using a keyboard shortcut to write text in any application
The status bar and dock display the dictation progress
Customize Every Step
Behind the simplicity lies a powerful process that you can customize to precisely write down your words
Note taken
Your words are saved as you speak, allowing you to focus on your thoughts. Search, edit, and share them in a calming interface.
Private.
The ultimate in privacy and convenience, dictop takes your spoken words and effortlessly transforms them into text, running OpenAI's Whisper fully on the device.
No telemetry. No data is sent to the cloud. Your words are yours and yours alone. Your usage is not tracked.
Long demo
Here is a walk-through of the app showing most of the features in action.
More accurate than you and me
A human-level understanding of speech with superhuman spelling and punctuation
Use Whisper on your Mac, or choose from OpenAI's Whisper, Deepgram or AssemblyAI services in the cloud.
If you want to use on-device transcription, Macs with Apple Silicon (M1 etc.) will give you the best performance because we rely on the GPU.
If you are using an Intel Mac and are not happy with the speed, the server-side transcription with better accuracy is available.
💻 macOS 14.0 or later is required.
You can choose from a variety of speech-to-text models, depending on your needs.
On-device models:
- Whisper Tiny
- Whisper Small
- Whisper Medium
- Whisper Large V2
- Whisper Large V3
Server-side models:
- OpenAI Whisper
- Deepgram
- AssemblyAl
UX: With the built-in dictation features, it was initially exciting to see the words appear and refine as you speak. However, in practice, it is distracting, and it restricts your movement, beause you have to stay on the text field. With dictop, you can dictate anytime and anywhere, regardless of your current activity. When you're ready to insert the text, only then navigate to a text area—or not. The transcript is automatically saved, allowing you to insert the text later by right-clicking on the dock icon or the status bar button.
Accuracy: You'll truly notice a difference in accuracy with Whisper and other models. Finally we have a speech-to-text model that we can use without the need for constant corrections.
Languages: We offer auto detection of the spoken language for people who use multiple languages per day, like many people who are browsing and interacting on the web.
For the server-side models, you can check them on their respective websites. Quoting OpenAI for Whisper :
It approaches human level robustness and accuracy on English speech recognition. Whisper's performance varies widely depending on the language. The figure below shows a performance breakdown of large-v3 and large-v2 models by language, using WERs (word error rates) or CER (character error rates, shown in Italic) evaluated on the Common Voice 15 and Fleurs datasets.
Choose automatic language detection if you switch between languages frequently, eliminating the need to adjust settings each time you speak. Alternatively, select a specific language if you primarily use just one. The following languages are supported by dictop:
- Afrikaans
- Albanian
- Amharic
- Arabic
- Armenian
- Assamese
- Azerbaijani
- Bashkir
- Basque
- Belarusian
- Bengali
- Bosnian
- Breton
- Bulgarian
- Catalan
- Chinese
- Croatian
- Czech
- Danish
- Dutch
- English
- Estonian
- Faroese
- Finnish
- French
- Galician
- Georgian
- German
- Greek
- Gujarati
- Haitian
- Hausa
- Hawaiian
- Hebrew
- Hindi
- Hungarian
- Icelandic
- Indonesian
- Italian
- Japanese
- Javanese
- Kannada
- Kazakh
- Khmer
- Korean
- Lao
- Latin
- Latvian
- Lingala
- Lithuanian
- Luxembourgish
- Macedonian
- Malagasy
- Malay
- Malayalam
- Maltese
- Maori
- Marathi
- Mongolian
- Myanmar
- Nepali
- Norwegian
- Nynorsk
- Occitan
- Pashto
- Persian
- Polish
- Portuguese
- Punjabi
- Romanian
- Russian
- Sanskrit
- Serbian
- Shona
- Sindhi
- Sinhala
- Slovak
- Slovenian
- Somali
- Spanish
- Sundanese
- Swahili
- Swedish
- Tagalog
- Tajik
- Tamil
- Tatar
- Telugu
- Thai
- Tibetan
- Turkish
- Turkmen
- Ukrainian
- Urdu
- Uzbek
- Vietnamese
- Welsh
- Yiddish
- Yoruba
Just like we localise the app and offer dictation in many languages, we also localise the name. The users using their Mac in different languages will see the version of the name in their language.
Yes. It's a Mac-assed Mac app.
We use Swift as the programming language, and AppKit and SwiftUI as the primary frameworks.
Not yet? Care for some text? :-]
- Seamless Voice-to-Text: Start recording your voice with a click or a keyboard shortcut, and dictop will begin transcribing your words into text as soon as you stop.
- Instant Text Insertion: The transcribed text is automatically inserted into the current document or text field, as if it was typed or pasted from the clipboard.
- Unobtrusive and Intuitive: The app runs discreetly in the background without any complex user interfaces, ready whenever needed.
- Productivity Booster: By harnessing the power of your voice, dictop lets you input text up to 2-3 times faster than traditional typing.
- Universal Utility: dictop is designed to benefit anyone who regularly types on their computer, no matter their profession or typing skill level.
- Simplistic Design: dictop focuses on one core task - voice to text conversion - and does it well. The app saves all transcriptions in a neatly organized format, accessible through a simple and intuitive interface.
- Multilingual Support: dictop transcribes 99 languages. It features an app UI translated into 11 languages, with plans to expand this coverage soon.
- Future Expansion: Plans are in place to bring the dictop experience to iOS devices, to provide the same seamless voice typing experience beyond macOS.