Ever since we released web results inside of Quick AI, I've basically not done any Google searches. Whenever I have a question, I just open up Raycast, press Tab, and ask it. If AI knows the answer, it will just give it to me, and if not, then it will search the web in the background and use that as context to then answer my question. That's just one of the ways that you can leverage the power of AI inside of Raycast, and honestly, that is the feature I'm using the most so far. We've actually shared a list on Twitter of how people have been using this feature, so I'm going to leave a link in the description below. If you're not following us on there yet, then maybe that's a good time to do so. But anyway, that's not all. Today, I'm going to show you four more ways that you can use AI inside of Raycast to make you more productive, more creative.

and honestly, just to blow your mind sometimes. Another way to use AI inside of Raycast is through the AI chat command. This is great when there is a little bit more back and forth needed to get the answer that you're after. It's also great because there's no context switching, so you can open up the chat from anywhere you are, and there's no need to be tabbing around until you finally find the chat. It's always just there, and it's ready to be used. The chats are also customizable, so you can give them a name, you can set the model, the level of creativity, and also system instructions, which I'll talk a little bit more later in this video. I personally use the AI chat command a lot when I need to do a lot of data structuring work. So for example, say I have a dataset in YAML and I wanted to convert it to JSON and also to help me with copywriting, where I'm trying to find the right words

or the right headline for something, and I find that the AI chat is a great command for that. Now you can make your chats even more powerful by relying on system instructions. That allows you to create chats that serve more like assistants. I've actually already made a video about this, but the idea is that you can set a system instruction on your chats, and you can define a set of rules and guidelines that you want your chat to follow. This is great for creating assistants because you can tell it exactly how you want the answer to be. I personally have an assistant for helping me translate things into Spanish, and also I have an assistant for helping me with anything related to front-end development. You can also use it in more creative ways. So let's say you want an assistant to help you generate bedtime stories for your kids. You can just give it the initial prompt, say a story about a unicorn and a rabbit, and the system instructions will contain the rest that is needed.

We're also working on a way that we can surface these AI chats to the root search of Raycast. That way, you can access them quicker, but you can also assign hotkeys to them. Then we have AI commands, and AI commands are a great way to automate some of the very common workflows. Out of the box, Raycast will give you a few commands such as fixed spelling and grammar, improved writing, change the tone to casual, but you can create custom commands that are more specific to your workflow. For example, I have a command called fill in the blank, and I use that often when I'm writing a sentence and I just can't think of the right word. That command will give me a few options of words that make sense in that context. Another command that I use often is one that transforms CSS style objects into vanilla CSS, and this one has saved me a lot of time as well. AI commands is such a great way to do these repetitive tasks for you.

You can assign hotkeys to them because they are already surfaced inside of Raycast root search. You can also use placeholders in them to make it more powerful. So you can use the selected text, as well as the clipboard text, and even an argument. You can also configure them to set the level of creativity and the model that you need. And the last tip is for all of the devs and people creating Raycast extensions, which is the AI API. Tom has recently rebuilt the reminders extension, and he's making use of the AI API. So let's get Tom in here to show us how he's done it. So we've rebuilt the reminders extension, and in this new extension, you have this command called quick add reminder. And basically, it uses Raycast AI to add natural language processing to the reminders that you write. Let me show you what I mean. I'm in the reminders native application, and I will add some reminders in it. For example, I'll remind myself to buy

groceries today at 6pm, and it's quite important. As you can see, it has been added for today, 6pm, with a high priority because I put important in the text. I can also add a recurring reminder. For example, clean the house every day. And so I now have a daily reminder to clean the house. I can also put a reminder in another list if I want. For example, fix a bug in the Raycast list, and I'll add some notes for this one. It has to do with authentication. Again, because I put the Raycast word in the text, it has understood that I was talking about the Raycast list. And so I now have this reminder here. Let's see what's happening in the code. What we do here is that we have a bunch of data that we use in a huge prompt. This is what we give to the AI. We ask it to return us a JSON structure. And so we write a lot of things, the JSON

structure, some rules to follow, the list names with their IDs, some examples to help out the AI. And then the AI returns us a result, the JSON data, and we parse over here. And then we can create the reminder. That's basically it. We have less than a hundred lines to add natural language processing to the way you want to add reminders in the app. And voila, that's it. Now, as you can see, AI is a huge part of Raycast. We're constantly working on ways to integrate AI into the product and to make most of our commands more powerful. We started building an AI playground, which is just a place where you can do almost everything with AI. We've also started exploring how we can use GPT vision inside of Raycast AI and how we can make it so you can interact with your entire file system in a super easy way. If you'd like to try out all of these Raycast AI features, but you haven't

subscribed to Raycast Pro yet, then check out the link in the description to try it out. And here's another video I think you might like, which are seven ideas on how you can make the most out of Quick Links.

5 Ways To Use Raycast AI in 2024

Scripts