Practical overview
When it fits
Use Whisper.rn when a React Native mobile app needs to integrate whisper.cpp-based transcription and the team can manage native setup, model files, and audio-format requirements.
Who it is for
- React Native developers adding speech-to-text to iOS or Android applications.
- Mobile teams choosing between Whisper and Parakeet transcription models.
- Developers building an app interface around local model-based audio transcription.
Documented uses
- Turn an audio file into transcription text within a React Native app.
- Initialize Whisper or Parakeet with a model file and pass audio for transcription.
- Add speech activity detection using the documented Silero VAD support.
Requirements & limits
The repository is public and displays an MIT license. The package is installed with npm; no account or pricing requirements are stated in the cited project documentation.
- The package uses prebuilt iOS and Android libraries by default, with documented options to build from source.
- Parakeet model files are large: the listed variants are approximately 356 MB to 1.26 GB.
- Parakeet accepts 16-bit PCM WAV files or raw mono 16 kHz signed 16-bit PCM data; it does not decode compressed audio formats.
- Core ML acceleration requires iOS 15 or later and separate Core ML model files.
- The experimental Hexagon NPU path is limited to qualifying Snapdragon devices; VAD and Parakeet contexts stay on the CPU.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Install Whisper.rn in a React Native project
Add the package and complete its documented native setup steps.
Before you start
- A React Native project targeting iOS or Android.
- Node package installation; iOS setup uses CocoaPods.
Steps
Install the package with the command below.
npm install whisper.rnFor iOS, run the command below after installation.
npx pod-installFor Android, add the documented
com.rnwhisperProGuard keep rule if ProGuard is enabled.
Expected outcome
The package and its documented platform setup steps are in place; transcription still requires application code, a model file, and audio input.
Guide limitations
- Expo projects must be prebuilt before using the library.
- This is an installation guide, not a verified build or transcription test.
Transcribe a WAV file with Parakeet
Initialize a Parakeet context from a downloaded model and request transcription of a WAV file.
Before you start
- A React Native project with whisper.rn installed.
- A downloaded Parakeet TDT 0.6B v3 GGUF model and a local WAV file.
- For Parakeet file input, WAV audio must contain 16-bit PCM.
Steps
Download a Parakeet TDT 0.6B v3 GGUF model.
Initialize a context with
initParakeet, providing the model's file path.Call
transcribewith the WAV file path and options such asmaxThreads.Read the result from the returned promise; release the context when finished.
Expected outcome
The documented API returns transcription output, segments, and an aborted-status field.
Guide limitations
- Compressed formats including MP3, AAC, and FLAC are not decoded.
- Raw audio passed to
transcribeData()must be mono 16 kHz signed 16-bit PCM. - This recipe describes the documented API flow and is not a human-run execution test.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 13 October 2026
- Original website title
- mybigday/whisper.rn · Official homepage title · retained in its source language