Transcribing isn’t fun at all. Good thing it’s something AI is actually good at. Aiko is an app for Mac, iPad, and iPhone that users Whisper—open-source technology created by OpenAI—to transcribe audio files. Aiko does not upload the file to the cloud to make the transcription; everything happens on your device. And it works fairly quickly, too: I was able to transcribe a half hour radio drama in just a few minutes.

The application works best on devices with Apple Silicon processors (Intel Macs are technically supported but are extremely slow at transcribing); my 2022 iPhone SE was significantly faster than my 2018 Intel MacBook Pro, which took around three minutes to transcribe 10 seconds of talking. If you have the right hardware, though, this application is just about perfect.

A screenshot of Aiko. The interface is clear—it just says


Credit: Justin Pot

To get started, you need to either point the application toward a file or start recording what you want to transcribe. You can add any audio or video file to the application, which will immediately get started on creating a transcription for you. The recording feature is mostly there for quick notes—the software advises you to record things using another application first if at all possible. The mobile version can grab audio from the Voice Memos app, which is a nice touch.

Three screenshots from the iPhone version of Aiko. The left shows a quick transcription; the center, the recording feature, which isn't much more than a microphone icon; the third, a transcription of the first episode of the classic Douglas Adams radio play


Credit: Justin Pot

The application will show you the text as the transcription happens, meaning you can start reading before the complete transcription is done. The application automatically detects the language being spoken, though you can set a different language in the settings if you prefer. You can even set the application to automatically translate non-English conversation into English, if you want.

It’s not a perfect application—there’s no way to indicate who is speaking when in the text, for example. It works quickly, though, and is completely free, so it’s hard to complain too much. This is going to be a go-to tool for me from now on.