Диктофон
A recorder that hears more
Recording, speech transcription, bird recognition and voice analysis all run right on the phone. Audio and texts stay on the device: no analytics, no ads, no trackers.
Offline first
Transcription, recognition and analysis work without internet. The models are downloaded once.
Data on the device
Audio is kept in the app's protected sandbox, which other programs cannot reach.
No trackers
No analytics, no ads, no third-party SDKs. All data stays on the device.
Clean sound
Two recording modes: "Speech" with noise reduction and "Nature" with an unprocessed source.
What the app can do
From recording and transcription to identifying birds by their calls.
Recording
Compact m4a/AAC format that plays on any device. With a live level meter while you record.
- 44.1 kHz - 128 kbps mono for speech
- Recording continues with the screen off (foreground service)
- Live waveform display and a speech indicator (VAD)
- "Nature" mode: 48 kHz, an unprocessed source for quiet and distant sounds
Speech transcription
Speech recognition works offline, right on the device.
- GigaAM (Sber, MIT), offline transcription of Russian speech
- Several offline models of different sizes: from light to accurate
- Long recordings are split into segments automatically
- A clickable transcript: tap a word to jump to that point in the audio
Bird recognition
Identifies the bird species by its call right on the device. Runs on the Google Perch model.
- Google Perch (Apache-2.0), thousands of bird species worldwide
- Fully offline: a TFLite neural network right on the phone
- A regional filter and calibration remove unlikely species
- Species names in Russian and Latin, with a confidence score
Voice analysis
Beyond the words, the app analyzes the voice itself - building a "passport" for it: color, note and profile.
- The color and note of the voice, from pitch, timbre and loudness
- Voice profile: type (baritone, soprano...), register, timbre
- Everything is computed on the device - the voice is not sent anywhere
Speech clarity
Speech scored on objective metrics: filler words, repetitions, pace and pauses. Handy for speakers and podcasters.
- An overall score of 0-100 with a color-coded rating
- Filler words, repetitions and speech pace (words/min)
- Long and filled pauses (uh, hmm)
- The report can be saved and shared as an image
Organizing
Hundreds of recordings at hand. Sorting, finding and sharing take a couple of taps.
- Folders, tags and favorites
- Sorting by date, length and name
- Search across transcript text, renaming and notes
- Playback speed 0.75x-2x and sharing of audio or text
- An optional backup to Yandex.Disk - you turn it on yourself
Privacy
Privacy by default. Nothing leaves the phone until you allow it yourself.
- Audio is kept in the app's private sandbox
- No analytics, no ads, no third-party SDKs
- Interface in Russian and English
Dark theme, a calm teal accent
A calm dark interface in an "audio-noir" style. The color of the icons hints at what each action does.
Basic is free, Pro is a one-time purchase
No subscriptions and no ads. All Pro features can be tried free for 7 days and then unlocked forever with a single purchase.
- Recording in two modes: "Speech" and "Nature"
- Folders, tags, favorites and search
- Notes, playback speed and sharing
- Everything is stored on the device
- Offline speech transcription on the GigaAM engine
- A clickable transcript and splitting of long recordings
- Bird recognition by call (Google Perch)
- Voice analysis: a "passport" - color, note and profile
- Speech clarity - a score based on objective metrics
- Text improvement with a neural network (SAGE)
What the app runs on
All the models are open and run right on the device.
Coming very soon to RuStore
No ads and no subscription. The Basic version is free, all Pro features are free for the first 7 days, then a one-time purchase of 499 ₽.
Coming soon Coming very soon toRuStore