Description of App
Whisper Notes: Offline speech-to-text powered by Whisper Large-V3-Turbo model, perfect for capturing daily voice memos and meeting notes. The app runs completely offline to ensure your voice recordings and privacy never leave your device.
No subscriptions, no in-app purchases, no ads—just a simple one-time payment for lifetime access.
■ Features
- Local voice recording & transcription
- Import and transcribe audio/video files, all offline
- Batch export or delete text and audio files
- Supports export of timestamped transcript
- Keyboard shortcuts for voice recording in any input field across apps
■ Why we made this
Sometimes I have a bizarre dream and want to record it when I wake up, but typing with sleepy eyes is tough. I needed a tool that could quickly capture and transcribe speech.
When friends send messages that need long replies while I'm out walking or eating, I want to respond quickly without sending voice messages.
■ FAQ
1. Does it support real-time transcription during recording?
Not currently, it can only record and then transcribe when finished.
2. Is there a maximum transcription length?
No, in theory you can record several hours of audio. However, since the transcription happens on your device, it will take some time. I've tested it with about 2 hour of audio and the transcription was accurate.
3. What is the transcription speed?
It depends on your device. On my M4 macOS, the transcription speed is about 1:12, which means two hours of audio takes about 10 minutes to transcribe.
4. What languages are supported?
English, 中文, Deutsch, Español, Русский, 한국어, Français, 日本語, Português, Türkçe, Polski, Català, Nederlands, Arabic, Svenska, Italiano, Bahasa Indonesia, Hindi, Suomi, Tiếng Việt, Hebrew, Українська, Greek, Bahasa Melayu, Čeština, Română, Dansk, Magyar, Tamil, Norsk, Thai, Urdu, Hrvatski, Bulgarian, Lietuvių, Latina, Māori, Malayalam, Cymraeg, Slovenčina, Telugu, Persian, Latviešu, Bengali, Српски, Azərbaycan, Slovenščina, Kannada, Eesti, Macedonian, Brezhoneg, Euskara, Íslenska, Armenian, Nepali, Mongolian, Bosanski, Kazakh, Shqip, Kiswahili, Galego, Marathi, Punjabi, Sinhala, Khmer, chiShona, Yorùbá, Soomaali, Afrikaans, Occitan, Georgian, Belarusian, Tajik, Sindhi, Gujarati, Amharic, Yiddish, Lao, Uzbek, Føroyskt, Kreyòl Ayisyen, Pashto, Turkmen, Nynorsk, Malti, Sanskrit, Luxembourgish, Myanmar, Tibetan, Tagalog, Malagasy, Assamese, Tatar, Hawaiian, Lingala, Hausa, Bashkir, Javanese, Sundanese
■ Contact
Comments
removing links?
The only complaint, and it's not a showstopper since you can copy just the texxt, is that you have links and time stamps by default. I couldn't find a way to hide those, but there is an option to copy just the text so that you can review it elsewhere. Worth the $6.99 for the rapid and offline transcription for sure!
Wow
This app sounds very promising. 👍
Developer update, and an answer to Scott
Hi everyone, I'm Ke, the developer of Whisper Notes. Thank you to the original submitter and to everyone who commented. A few details in the entry above have changed since July, so here is the current information.
Price: $7.99 in the US (prices vary by country), one time, no subscription or in-app purchases. The same purchase also unlocks the Mac App Store version, so you don't pay again for it.
Speech engines (everything runs on the device; recordings and transcripts are not uploaded):
On iPhone and iPad:
- Parakeet (default). Moondream's Parakeet Ultra, post-trained from NVIDIA Parakeet V3
- Whisper Small
- SenseVoice
- Apple Speech, on iOS 26 (supported devices)
On Mac (the Mac App Store version, included in the same purchase; needs Apple silicon):
- Parakeet (default)
- Whisper Large V3 Turbo
- Whisper Small
- SenseVoice
- Qwen3-ASR (beta, macOS 15 or later)
- Apple Speech, on macOS 26
VoiceOver details:
- Magic Tap (two-finger double tap) starts and stops a recording from the main screen and the recording screen. VoiceOver says "Recording started", and "Recording stopped. Transcribing." when you stop.
- VoiceOver announces when a transcription finishes, fails, or finds no speech, so you don't have to check the screen.
- In a transcript, each sentence can be activated to start playback from that point. Playback speed and playback position are labelled.
@Scott, about timestamps: if you only want the words, use Copy Text, or Share or Export and choose TXT. The versions with timestamps are separate items (Copy with Timestamps, Export TXT with Timestamps), so the plain text never includes them.
If anything in the app is hard to reach with VoiceOver, Braille or Zoom, please tell me at support@whispernotes.app. Specific steps help me fix it quickly.