Transcription
In this video, we will look at super mode. This is what you use when you want dictation that adapts to whatever app you're in without having to stop and think about whether the results will come out right. It takes advantage of AI to give you more accuracy and a flexibility that is not available in the other modes we have covered so far.
Let's go to the modes tab in Super Whisper and create a new one. I'll pick super from the top of the list. If you click the gear icon to open the sidebar, this looks like any other mode, but there's something different happening behind the scenes. This mode is great because it understands what you want to say, not just what you literally dictate.
I'll switch to super and open a new note. Hey, John, let's meet on Friday afternoon. Actually, no, let's do Monday morning. As you'll see, this mode catches when you correct yourself mid-sentence. It completely removed my mistake.
But there's another thing that makes this mode even smarter. It captures three types of context to help the AI improve the results. The first is application context. It grabs names, vocabulary, and content from your active window. So, the AI gets the spelling right. If you're in an input field, it will even grab the content from there. Say I'm writing about a fictional product called Hush Pocket. And there's also the name Rebecca spelled with K and H. Let me dictate something for Rebecca. Using the Hush Pocket made all the difference. With super mode, I don't even have to add any of those words to my Super Whisper vocabulary. It will just intelligently detect the correct spelling from what I already have in my window. Super Modes application context is all about getting good results straight out of the box. The model is not acting blind anymore and you don't even need to worry about it. You can just let it do its thing.
But you can also choose to be more proactive with the AI using the second context type available selected text. Here you can take it beyond the automatic spelling fixes or selfcorrections. And you can actually start giving specific instructions to transform text. Let me start dictating an email. Hey Sarah, new line. It was great meeting you yesterday. There's still a few things that I need from you. Send me your tax forms, a link to your portfolio, and also I need references that you can give me. Once I have that, I will send it to the business department, and we can talk about what's next. Best, Robert, if you noticed, I dictated the new line instruction to give a quick hint to the AI. It knows that I'm on my mail app. It recognized that I wanted the greeting in a separate line. And with that, it understood my intention for the rest of the dictation.
But now I will select this text and dictate a short command. Make this sound more professional. Use bullet points where necessary and also add a line saying that we are excited to have her join the team. This way I can get a quick improvement without having to copy and paste into some other AI app. And the cool thing is that since we got a recording window which automatically detects if we are in an input field or if it should hold the result for us. Well, we can also use this in PDFs, documents, websites and more. Here I am on Super Whisperer documentation. I will select some text and say, "Hey, translate this into French for me." This can be super handy because everything is just one command away.
Aside from app context or selected text context, there's also clipboard context. This is perfect if you need text from another window or application. Whatever text you copy will also inform the spelling of names and terms in your dictation. The important thing to note here is that selected text is captured the moment you start dictating, but clipboard context is captured a couple of seconds before or during the recording.
For example, let's say that I'm sending a message to a user explaining something a bit technical about a JSON file. Hey, William, in the JSON file, application context enabled being true or false just means that information is being passed on to the AI to improve the results. And language model name is just the actual model that is being used for your request. I suggest switching to a more capable model for better results. So, I started my dictation with nothing selected. Then I went to copy something and I came back to my note for the result to be inserted. As you can see, those technical terms were spelled correctly.
Let me show you something else with super mode. When using the classic recording window in the bottom left corner, you can see the application from which context will be captured. If I switch to something else, this will also change. You just need to make sure you are in the right window when you finish dictating. If I copy something in the middle of my dictation, I get this indicator showing that clipboard context was also captured. If on the other hand, I start dictating with text already selected. I will have that indicator from the very beginning. This is just there to inform you that everything is working properly.
So, the possibilities really open up here. You get a mode that's flexible enough to improve your results intelligently. And it can also do some of what the other modes can do if you start transforming text via commands. But it's also good to know that this is still meant to be used for text formatting and not necessarily something for generating long form content. That means that if I dictate something like draft a short email to Josh with a reminder for our meeting tomorrow afternoon, it will just transcribe that AI will not actually draft my email. In other words, it's great for transforming text you already have and for giving you accurate results depending on your current app, but don't really like a full-blown assistant. You can still get that functionality with Super Whisperer, but we will learn more about that when we start exploring custom modes.
Also, since this mode relies on the intelligence of AI, the model you are using matters. Remember, you can change that in your modes tab by opening the sidebar. Another thing to consider is that since this mode pulls more context from your active window, processing could potentially take a bit longer. Super mode is flexible and works like an all-in-one solution, but the other modes are still great for consistent, straightforward results. For now, go ahead and test this out. Create a super mode and try it with different models so you can see what works best for you.