Tips for Using the AI Coding Editor Cursor (Summary)

00:00 Congrats for making it to the end of this Code Conversation. Here’s a quick summary of all the things that you’ve covered watching through this course: You saw how to install and set up Cursor and switch on privacy mode.

00:12 You got to know Cursor’s interface, both the agent view and the editor view. Then you worked with Plan mode and Agent mode. And actually, let me pause here for a second and take a little step to the side because Cursor has more to offer than just Plan mode and Agent mode.

00:29 There are a couple of different interaction modes that you didn’t cover in this course at all. You did look into plan mode, which can help you to organize and streamline coding strategies and make architectural decisions before starting to build anything.

00:42 And then also the agent mode, which does this AI-driven code generation that you’ve worked with a lot in this course. But there’s more: Cursor also offers an Ask mode that you may have seen when I was selecting between different modes, that essentially gives you a context-aware chat interface similar to what you would get if you are just using Chat GPT on the site.

01:04 But in this case, it’s context-aware, so you can reference files and it’ll know how to pick the context out of your code project. But it won’t do anything: it won’t create a plan and it also won’t execute any code.

01:17 It will give you answers that are related to your project and also give you suggestions, just not implement them. So this is also a useful mode that you may want to work with.

01:28 Then we have debug mode, which is a dedicated debugging workflow where the agent generates multiple hypotheses for what may cause a bug and then instruments your code with logging.

01:39 Then it actually hands control back to you, asks you to reproduce the bug, and then uses the runtime logs that you collect by reproducing the bug to exactly pinpoint the actual root cause.

01:51 With that info, it proposes a small targeted fix and finally asks you to verify the bug is fixed before removing the logging it wrote. So a pretty cool feature to try out.

02:01 Then remember the built-in browser I showed you? Well, you can do more with it than just view pages and select elements. The visual editor in Cursor’s Browser lets you edit your web UI visually. Think drag and drop elements, tweak styles with button controls, and point to specific elements and prompt changes directly on the rendered page.

02:21 The cool thing is that throughout all of this, the agent updates the underlying code, keeping your visual updates in sync with your code base.

02:29 Cursor also gives you more. It gives the option to do inline edits. So that is if you’re inside of a file and you press– on macOS, it’s Command + K or Control + K otherwise, then it opens up a little chat interface that you can use to implement edits directly related to where you are in the file.

02:49 And there’s also Tab Complete, which allows you to jump through AI-assisted edits and perform smart code completion across a file. Looks like that you just keep pressing Tab to accept suggestions that Cursor provides for you, again, within the context of your project.

03:07 So those are also cool features that you may want to take a look at if you start working with Cursor more. Back to our summary.

03:15 So while exploring agent mode, you also saw that you can select different models and that you can even select multiple models and run them for the same task and then decide which of the outcomes you like best.

03:29 You’ve seen me use voice input and also one voice-submit. And we’ve also touched upon how Cursor uses version control, how you can use version control with Cursor, and also resolve the very trivial merge conflict, but still to give an idea of how those workflows integrate with Cursor. You got to see the very useful built-in browser and how it renders a page that you built and how you can select elements and add them to the chat to ask questions or ask an agent to apply changes to that specific element, and how this can be very useful also to help you identifying and fixing bugs. I’ve also switched off control between agents and myself as driver multiple times.

04:13 So you’ve seen how you can do that at any time. Finally, I learned a new Bash command, and with that I also wanted to highlight that the process of developing in an AI-assisted coding environment can also help you to learn new things.

04:27 So it’s not just about doing less, but it’s also giving you a chance to encounter new ways of doing things and then also integrate them into your personal tool belt.

04:38 And with that, I want to give you a couple more additional resources. If this sounded interesting, but you’re not that interested in working with an IDE, then you can go ahead and check out the code conversation on getting started with Claude Code, which is a command-line interface for AI-assisted coding.

04:56 And the same for Google’s Gemini CLI. So both of these are also tools for AI-assisted coding that you can access directly through your command-line interface.

05:08 So it’s a bit different than an IDE like Cursor, but it can also give you a way to work with AI models in your coding workflow.

05:19 And finally, I want to mention the AI coding tools reference area at Real Python, where you can find more information and other tools that you may or may not have heard of and read some quick definitions on what they are and what you can use them for.

05:34 Thanks for watching and keep exploring. See you around.