I needed to transcribe some audio and only found web apps that were very limited and free or expensive. None very private. Also the pricing is not great because often there is only a monthly or anual subscription which is not good price per usage when I need the occasional transcription.
Free tier also have low file size or minutes allowed. I had a 4 hour file.
My first version is a CLI tool that offers pretty much the same features, just built in JS and runs in Node. The biggest difference is that it uses ffmpeg directly instead of ffmeg.wasm.
Soon I wanted to transcribe a file from my phone and the web app idea appeared. The main difference from the “competition” is that this web app runs fully on the client side with local storage. It doesn’t send any data or keys to any other service than the OpenAI API. The upload section of this web app will just load the file to transcribe into the context of the web app so it can manipulate the file, otherwise it doesn’t have access to the file system.
➡️ App
Next.js 15, React 19, Tailwind CSS v4, FFmpeg.wasm, OpenAI API
Key Features & Technical Achievements
🎯 Intelligent Audio Processing