Build a Personal Assistant with GitHub Copilot SDK + Copilot CLI

I want to show you how you can build your own personal AI assistant on the Copilot SDK, how easy it is to do that. So, the first thing that we're going to do here is we're going to get Copilot CLI to learn everything that it can about the Copilot SDK. And we're going to use this new research command that's built into the CLI, and then we're just going to have it research github/copilot SDK. And that is the repository for the Copilot SDK. It contains all the code, all of the docs. Essentially, I want the CLI to know everything it can possibly know about the Copilot SDK so that it doesn't try to do things that don't exist in the SDK or write code that won't work. Okay, so here's our full report right here. If we click on this, here it is. You can see it's quite lengthy. Now, we're going to want to make a plan on this knowledge of the SDK. So, I'm going to shift tab into plan mode here, and I'm going to say, "I want to build a personal assistant on the Copilot SDK that I can communicate with over Telegram." And then, I'm going to use the at sign, and I'm going to specifically mention our research here, the markdown file. And then, I'm going to go ahead and send it. Okay, so our plan is ready, and I'm going to go ahead and choose number one, which is accept plan and build on autopilot plus fleet. Fleet is where it deploys a bunch of sub-agents to take care of things, and it tries to continue working until it gets all the way to the end. So, this is going to take a minute to complete. We'll fast-forward and come back and talk about the results. So, I will see you in just a second. Okay, it's done, and you can see there's a lot more files in our project now, especially in the source folder here. Okay, so update built a fully featured Telegram personal assistant on the GitHub Copilot SDK. Awesome. Now, the question is, how do we run it? Okay, so to run, copy env.example to env. Okay, so here's env.example, so I think we can just rename this, right, to env. Perfect. And then, add your Telegram bot token, ensure Copilot auth login is done, and then run npm run dev. Okay, so let's go ahead and get this Telegram bot token. Let's ask. Now, before we do that, I'm going to switch to a smaller model here, like Haiku. Um and I'm just going to say, "Let's get switch out of autopilot. Where can I get the Telegram bot token?" I actually know how to do this, but let's just assume for a second we have no idea how to set up a Telegram bot. I want the CLI to walk us through that. Okay, so we need to open Telegram and search for BotFather, start a new conversation, create a new bot, follow the prompts. Okay, and then we need to paste in our tokens into this file here, right? And it looks like our Telegram bot token is what we want, and we want to paste that into our env file there. Okay. So, in Telegram, if we just search search for BotFather, you'll find it right here. It's very popular. And then, what can this bot do? And then, let's start a chat. And now, we want to do new bot. All right, a new bot. What are we going to call it? We can call it anything. Let's just call it personal assistant. And then, we're going to have to choose a unique name for it. And so, I'm going to call this Burks. Let's just call this Burks personal assistant demo bot. And then, it has to end with bot or _bot, as it shows in the example there. And then, we get this key that we need. So, we're going to copy this right here. So, I'm going to control-C copy that, and then paste that in right there as per the instructions. And then, we can click on this, and that will actually add the bot. So, here's the bot, and we can start it. Now, our personal assistant isn't running. So, we need to start the personal assistant. So, let's go back here, and it looks like the default model the default model, you know, I'm going to recommend that we use a Sonnet 546. It's a very good model, Sonnet 46. And okay, so then, let's come back here, and we want to start this, npm run dev. Starting the personal assistant Telegram bot starting. Send start to begin chatting. Okay, so in theory, if we come back to Telegram here, we should be able to say, "Hello." Okay, and we got a thinking response right away, which I think came from here. Yes. Create a session. Okay. There it is. So, just like that, we're up and running with Telegram, and we have a bot. "I'm here. I'm the Copilot CLI, your terminal assistant. I can help you." And of course, it thinks it's a coding assistant. So, we would need to go back and work on this. Thanks for watching this video. I really like these personal assistants, and I don't think they're overhyped. I think they're actually really cool. Go build your own. And as always, happy coding.