Getting accurate speech-to-text has always been a pain in the ass. A decade ago, it would cost you a $199 app purchase—and the result would still require a lot of editing to make it accurate. But this is where we start to see tangible benefits of the AI revolution, where the learning models can do things incredibly quickly, and without using a ton of resources. While OpenAI’s Whisper technology is shockingly good at turning speech into text, you need to be a developer or a technician to actually make the most of it. But now, a developer has done the heavy lifting of turning this technology into a delightful little Mac app: MacWhisper.
You can use MacWhisper to directly record audio to be transcribed into text (with time stamps). However, things become a lot more interesting when you import an audio file or a video file—the app can quickly generate accurate transcription, time-stamped down to the millisecond.
Once the text is generated, you can go through and edit it. The Reader feature can show you all the text together in a document preview; you can then copy the transcript. Click the Share button, and you’ll be able to download the entire transcript as an SRT file (this is where those timestamps really come in handy).
You can download the MacWhisper app for free via Gumroad—put a “0” (that’s a zero, not the letter o) in the payment field to get it for free. Currently, the app requires macOS Monterey or Ventura, and it’s recommended to use an Apple Silicon Mac for fastest results. The free version will be enough for most people, but if you plan to use this in a professional setting, we suggest you spring for the €9 Pro version. This gives you access to the 3GB large data model, which improves the accuracy.