Exploring the OpenCode Interface

00:00 It’s almost time to get started implementing features with OpenCode. The final thing left to do is for you to get more or less acquainted with the interface that you’re going to be using, which is the OpenCode TUI.

00:11 Now, you might see already on the screen that there’s a keyboard shortcut that you can try. Ctrl+p will open a modal screen that contains a list of all the commands that you can run within the OpenCode application.

00:26 Now, remember that OpenCode is an open-source coding agent. It sits on top of LLMs and provides a number of functionalities that make you more productive and make it easier to work with LLMs.

00:39 For example, if you find the section that talks about the session, you see commands that allow you to switch sessions, to create new sessions, to move sessions to other project directories. A session here is a quote unquote conversation, and so you can see that you have commands that allow you to manage your conversations with the models. Explore this list, use the keyboard arrows to go up and down. If you want to run a specific command, press Enter to run it, and once you’re done exploring, you can press Esc to exit the command palette.

01:15 Now, another way of running commands is through the forward slash. If you type a forward slash, you will see a list of commands. Again, you can use the arrow keys to go up and down.

01:26 Alternatively, if you already know the command you want to run, you can just type it. For example, /quit or /exit, you can exit OpenCode.

01:36 Open it again by typing opencode.

01:40 Another command that’s useful, that’s not immediately visible on screen, but that you can find if you press ctrl+p to open the command palette, and if you scroll down to the agent section, you will find a keyboard shortcut, Variant cycle, ctrl+t.

01:57 This allows you to cycle through the different model variants that are available, which control the reasoning level. For example, for Gemini 2.5 Flash, as you press ctrl+t, you will see right in the middle of the screen the words high,

02:16 max, and if you press ctrl+t again, nothing. The nothing is the invisible setting, which is the default, which allows the model to pick the reasoning level that it determines to be the most appropriate for each task.

02:32 That’s why it’s the default. This is a sensible value to go with. Different models, different providers, they may have more levels, fewer levels, and the levels might have different names.

02:43 That’s perfectly fine and it’s normal. So don’t worry about that, and make sure to press ctrl+t until you leave it in the default, which is the invisible.

02:53 And finally, there’s another keyboard shortcut tip on the screen, which is if you press Tab, you’re going to cycle through your different agents. And for that, you’re going to want to look at the word Build that’s on screen.

03:08 So if you press Tab, that’s going to switch to Plan. And if you press Tab again, it switches back to Build. So these Build and Plan modes, or these Build and Plan agents, are going to be very important in the next lesson when you start planning your new feature.