MacWhisper: This Mac App Saves Me Hours By Typing What I Speak
Reviews • June 17, 2024
Good typing speed is necessary if you work with computers and documents. Depending on your job, a high typing speed of 80 WPM (Words per Minute) or higher can make your life easier.
I say this from my experience as a content writer for ten years.
Some job positions may even include a minimum typing speed under their requirements.
But what if you could type faster without typing?
What if you could say what you want to write? Like that scene from Her.
We do not yet have Scarlet Johanson’s voice, but there is a tool that can change how you type: MacWhisper.
TheSweetBits’ highlights
- MacWhisper is a transcription app for Mac that can significantly reduce writing time by converting spoken words to text.
- It offers multiple transcription models with varying accuracy, speed, and storage requirements.
- The free version provides basic functionality, while the paid Pro version unlocks features like global mode and additional language models.
Why MacWhisper?
Imagine this scenario: you want to submit a term paper tomorrow with 2000 words to be written.
How would you go about it?
If you plan a structure and type the entire thing, you can only pull it off if you have a high typing speed.
But this seems possible with a tool that can accurately transcribe your words. You can speak your mind and construct sentences as you do it. A few hours and some editing later, you will have that 2000-word term paper ready.
Well, this is the core idea of MacWhisper, but it does more than help you submit the term paper on time.
In this age of artificial intelligence, it makes little sense to stick to your keyboards, right?
What is MacWhisper?
MacWhisper is one of the most popular audio transcription apps available on the Mac. This application uses the Whisper model from OpenAI to transcribe audio in more than 100 languages.
MacWhisper can upload your audio via Whisper API or use a locally available transcription model, depending on the feature.
Here’s how it works:
You can input an audio recording from multiple sources.
- You can speak what you want to write.
- Upload an audio recording that you already have.
- You can also upload a podcast episode if you wish.
MacWhisper will transcribe the audio and provide you with editable text. This used to be a dream for me a few years back. But, with apps like MacWhisper, it is a reality.
For one, I did not WRITE this review. I spoke to MacWhisper, it was transcribed, and I did some editing.
In this review, I will share all the great, good, and not-so-good things about MacWhisper and how it can help you write things faster without compromising quality.
Getting started with MacWhisper
Setting up MacWhisper and using it for the first time is easy.
You can download the application for free from the official website. Once the app loads up, you will be asked to download a transcription model.
Some of these models might not be available depending on whether you have a paid version. You can choose one of the suitable models based on quality, speed and size.
Once MacWhisper has downloaded the model, you are good to go.
Different Transcription Modes
As I said, there are different ways to convert your speech into text with impressive accuracy.
The first method is to open an existing audio/video file. This could be an MP3, M4A, OGG, AAC, or MOV file on your Mac.
The second method is to do a new recording. This way, MacWhisper will record speech from your side and transcribe the content immediately.
You can also paste the URL of a YouTube video, audio or video file URL. MacWhisper will download it for you and transcribe the audio.
The Global mode lets you convert speech to text and paste it instantly anywhere on your Mac.
The Transcribe Podcast feature is also available in the beta stage, but you can try it out.
There are also additional options, such as batch transcription and cloud transcription.
Batch transcription works excellent if you want to transcribe data from multiple audio files.
On the other hand, cloud transcription is reserved for those who do not have a powerful Mac. In that case, you can use the Whisper API to get the job done, but the API will cost you money, and you have to set it up.
MacWhisper: the great
Here are some sure-shot reasons to get MacWhisper.
Seamless
With a faster internet connection, MacWhisper does not take more than a few minutes to set up. Once you have set up the hotkeys, it is seamless to transcribe audio from multiple sources.
I am a big fan of the global model because that is the most accessible option. In this mode, you can launch the transcription utility wherever you are on Mac. It will record the content and copy the text to the clipboard. You will also have an option to make some edits to the transcript.
This feature is excellent when writing on something like Microsoft Word or Notion. I do not use this particular option, but you can have Mac Whisperer automatically paste the transcript wherever your cursor is.
Overall, Mac Whisperer is an install-and-forget thing for the most part.
Multiple Transcription Models
Everyone does not need the same level of accuracy or speed when it comes to transcription. So that you can select a suitable transcription model, it is a great choice here. And I will suggest a few things.
Note:
For the record, I have tested all these features on a MacBook Air M2 running macOS Sonoma. Though I have tried transcription in other languages, I will use English as the source for this review.
You can always choose between multi-language models and models optimized for low RAM devices.
When transcribing English, you can choose the Base (English) model or the Small (English) model. Both of these are available in the free version. These two models are accurate enough for most tasks. However, the transcription struggles with proper nouns or brand names.
If your documents contain proper nouns, you may need something more powerful. Thus, if you want a good balance between quality, speed and accuracy, you should get Medium (English). This model will take up 1.5 GB of space on your Mac and takes a little toll on the processor when running at full speed.
MacWhisper also lets you download all available models, and you can shift between these models as you need.
Flexible Design
I am also a great fan of the flexibility that each MacWhisper option provides.
For instance, you can decide what happens after the recording is transcribed while setting up the global mode.
In my case, I always set it to copy the content to the clipboard. But there are multiple options.
For instance, if you want to transcribe a YouTube video, you can choose the quality of the content you download. You can also use the MacWhisper menu bar app if you need a quick way to transcribe something.
When it comes to transcribed content, you can ask MacWhisper to ignore segments. For instance, you can set the application to ignore blank audio, silences or crunching voices. You can customize these aspects based on your recording environment.
Video Player
The latest MacWhisper 8 introduces a powerful video player that significantly enhances the transcription experience. It seamlessly displays translated text as separate subtitles, allowing you to follow along with the video without constantly switching between the transcript and the video, making the editing process much smoother.
If you only need the transcribed content and not the video itself, MacWhisper offers the option to download just the audio portion of a YouTube video. This feature not only saves storage space but also lets you focus on the transcription without the distraction of video content.
MacWhisper: the good
Now, we will discuss a few areas where MacWhisper works just right.
Performance
I have a somewhat confusing take on the performance of these models and how optimized they are.
I used the Medium (English) model during my testing, which is not the most extensive but is just right for accuracy. When using the Global feature, MacWhisper took a lot of time to process the transcript.
I had a similar experience while transcribing a YouTube video.
But I do understand that I am running an M2 Mac, and things could be better on an M3 chip.
However, performance is not an issue if you are working with the other two models I recommended. MacWhisper will make sure that you get accurate transcription in a few seconds.
Customization
MacWhisper leaves decent room for customization and control over all these available modes.
For instance, you can use the Find & Replace feature to deal with frequent/potential errors. Similarly, you can ask MacWhisper to ignore specific segments. Lastly, you also have an option to add speakers to your transcripts.
It’s also great that you can adjust the UI. MacWhisper lets you choose between the transcript and segments mode. And you can always change the font size and add timestamps to the transcript.
MacWhisper: the not-so-great
These are not dealbreakers, I promise.
Cloud Transcription
I understand the reason behind this, but things can be difficult if you have a Mac with 8GB of RAM, especially when multitasking. It would have been nice if Mac Whisper offered the option for cloud transcription.
If you want to use the Whisper model on the cloud, you have to get a separate API key, and that requires an additional setup.
Or, MacWhisper could introduce a subscription plan offering some Whisper credits for cloud transcription.
Apart from this, I love everything about MacWhisper.
Accuracy
The accuracy aspect of MacWhisper is dynamic.
With the small and medium models, MacWhisper had almost 99% accuracy while transcribing English. But, as I said, the Small model may not get brand names/capitalization correctly.
However, the accuracy levels may differ based on the language and the available model.
There are also some languages where AI training data is not widely available, thus affecting the transcription option.
I have used MacWhisper’s transcription tool to write blog posts and research content. The research content made use of some complex vocabulary, and the accuracy of MacWhisper was great.
Pricing and availability
MacWhisper is free, but the free version has some restrictions.
- You do not have access to the Menu-bar app or Global mode.
- However, you can record audio and get it tagged right.
- Similarly, you can’t access batch transcription.
- Only a few models are available.
- Some customization features are unavailable.
But you can go for Mac Whisper Pro if you need these features.
The MacWhisper Pro license is available for 30 Euros ($32). But you can save some money if you want to purchase multiple licenses. With this license, you can use the power of transcription even offline. MacWhisper offers unparalleled privacy and security as your data is not sent to any servers.
The good part is that you do not need the Pro version unless you require global mode and additional language models. It is also a good time to mention that you should have a Mac with at least 16GB of RAM to run the medium and pro versions of Whisper.
Overall, for what it offers, MacWhisper has a competitive pricing structure.
TheSweetBits’ final thoughts
As you might have understood, I became a fan of MacWhisper in no time.
MacWhisper is a worthy investment for people who must type a lot daily. These days, I spend half of the time than I used to for writing an article. Why? Speaking the words and editing them is much faster than thinking and typing.
Because you are downloading all the models to your device, there are also no privacy concerns. So, MacWhisper is a better choice for confidential material than the other transcription tools I have used.
Overall, we suggest that everyone check out what MacWhisper can do for you, it will make your life a little bit easier.
