WhisperJournal

Privacy-first voice journal app built using React Native using whisper.cpp through whisper.rn for local transcription of voice recordings.

Features:

• Autodetect language, for multi-language support
• Translate non-English audio to an English transcript
• Integrated noise reduction using AFFTDN in FFMPEG
• Model selection
• Full privacy thanks to local inference; the only network calls are to download transcription models, not even analytics/telemetry
• Android and iOS
• Dark mode
• Open-source

Dev environment setup

Tested on MacOS 13.4.1 running on Apple Silicon (M2 chip), using:

• Node 18.17.0
• Java 11 (openjdk 11.0.20 2023-07-18 LTS)
• XCode 13.4.1

Step 0: Set up local env for cross-platform development

Follow the instructions under React Native - Environment Setup for your platform under the "React Native CLI Quickstart" guide for both target OSes until the "Creating a new application" step.

Step 1: Install all the things

Assuming you've already set up all the other things, yarn should ship as part of node; as a result, you can just run yarn and everything should install.

NOTE: You might have to run corepack enable to activate yarn, if you just installed node.

git clone git@github.com:kaizoco/WhisperJournal.git
cd WhisperJournal
yarn

Step 2: Build app

yarn start

yarn android
# or for ios
yarn ios

About Kai

I make cool things, check out my personal website! You can support future development of this (and other projects) from my Ko-Fi page, thanks for taking a look <3