WEBVTT 00:00.280 --> 00:07.720 If you are using CPS or want to learn how to use them, this new MCP toolkit developed by Docker will 00:07.720 --> 00:10.320 save you a lot of time and make it much easier. 00:10.640 --> 00:11.800 It's really impressive. 00:12.000 --> 00:16.920 And in this video I'm going to show you exactly how this works and how to use it. 00:16.960 --> 00:19.200 MCP stands for Model Context Protocol. 00:19.560 --> 00:27.120 It's an open standard created by anthropic that helps AI models talk to real world tools like calendars, 00:27.480 --> 00:30.200 CRM, emails, or your local files. 00:30.560 --> 00:37.920 So basically, instead of building custom connectors, every time you want to give you a access to something, 00:38.600 --> 00:42.680 you can use MCP as a shared language between models and tools. 00:43.080 --> 00:45.680 So MCP solves that by acting like a bridge. 00:46.200 --> 00:53.840 Once you set up your model, can read emails, write code, update CRM, or fetch live data all securely, 00:53.840 --> 00:55.880 and you can do all those things from one app. 00:56.120 --> 01:03.200 Now, while MCP is a powerful concept and you can already use it for a lot of useful things and use 01:03.200 --> 01:03.760 cases. 01:04.040 --> 01:07.960 It's still early days, so the ecosystem is kind of messy right now. 01:08.520 --> 01:13.080 There are a lot of apps and tools to choose from, but not much guidance. 01:13.320 --> 01:20.560 Many MCP servers are open source, so you can find them on GitHub, but they require manual setup. 01:20.720 --> 01:23.200 So normally you would need to go to GitHub. 01:23.840 --> 01:29.440 Find the right repository for the MCP you want to use, and then set everything yourself. 01:29.680 --> 01:36.280 And that includes downloading the code, managing dependencies, and in some cases even hosting the 01:36.280 --> 01:44.680 service on your own machine and connecting llms to tools like Cloud Desktop App or Visual Studio Code, 01:44.880 --> 01:48.320 or even a simple API takes a lot of manual setup. 01:48.760 --> 01:53.760 So you need to know how to set up environment variables, credentials, permissions, and so on. 01:54.240 --> 02:00.480 And because of that, if you are just starting out and you don't have any programming background, It's 02:00.480 --> 02:03.520 tough to know which MCP servers you can trust. 02:03.560 --> 02:10.160 A lot of people end up relying on random blog posts, GitHub threads, or community lists and hope the 02:10.160 --> 02:14.440 setup will work and that nothing shady is inside the repo. 02:15.000 --> 02:20.320 Yes, so the setup process is usually pretty manual, so you often need to clone the code, install 02:20.320 --> 02:24.400 dependencies, and figure out how to run these services locally. 02:25.000 --> 02:31.080 And when you are running tools that can access your machine, that can become a problem at some point 02:31.080 --> 02:33.600 because security is one of the biggest concerns here. 02:33.960 --> 02:40.440 So many of these tools have wide access to your system, and credentials are sometimes passed in plain 02:40.440 --> 02:41.080 text. 02:41.240 --> 02:49.160 So that means it's possible for sensitive information like environment variables, API keys to be exposed. 02:49.680 --> 02:56.040 And if you are working in a team, it's even harder to use because most setups don't come with basic 02:56.040 --> 02:58.160 protections like permissions controls. 02:58.480 --> 03:03.360 And that's a big reason why adoption has been slow, especially on the enterprise level. 03:03.760 --> 03:10.560 Also, every time you try to install one of these MSPs and it fails, it often still sends API requests. 03:10.760 --> 03:15.840 So at the end, you might get a lot of errors and spend money for each API call. 03:16.200 --> 03:18.320 So yes, it can be really frustrating. 03:18.720 --> 03:20.240 And that's where Docker comes in. 03:20.520 --> 03:27.200 Just to give you a bit of context, Docker is a really useful tool that lets developers run apps in 03:27.200 --> 03:30.520 isolated environments called containers. 03:31.000 --> 03:33.680 So it's widely used in programming and DevOps. 03:33.800 --> 03:38.560 And honestly, I could make a whole separate course just on Docker itself. 03:38.880 --> 03:46.080 So instead of manually setting up your own server and figuring out how to set up all this technical 03:46.080 --> 03:52.760 stuff, you can now launch verified MCP servers in separate containers with just a couple of clicks 03:53.200 --> 03:54.320 from Docker Desktop. 03:54.600 --> 04:01.680 Docker desktop helps you run these MCP Servers safely without messing with your system setup. 04:02.280 --> 04:08.120 So what that means is you can access all these MCP tools created by developers. 04:08.760 --> 04:13.440 And Docker released this toolkit where you can see a whole list of MCP servers. 04:13.960 --> 04:17.480 And these are all secured verified servers that you can access. 04:18.680 --> 04:25.560 And you can connect these MCC to apps like cloud desktop app cursor, Visual Studio Code. 04:26.800 --> 04:31.360 So to the MCP clients and automate a lot of tasks. 04:31.760 --> 04:35.760 So I will show you exactly how to set this up and how it works step by step. 04:36.080 --> 04:42.320 I will go through how to get started with Docker MCP toolkit, how to install it and launch your first 04:42.320 --> 04:45.840 server, and how to use some of the tools on real examples. 04:46.560 --> 04:47.680 So let's get started. 04:47.720 --> 04:50.800 Obviously to start you will need Docker Desktop. 04:51.480 --> 04:55.800 So go to docker.com and you should land on this page. 04:56.360 --> 05:01.760 So simply download Docker Desktop on your operating system and install it. 05:02.640 --> 05:04.720 Now, once you're done, please go ahead and open it. 05:05.280 --> 05:13.880 Now, before just a few days ago, you had to open extensions in the left hand menu and search for MCP 05:13.880 --> 05:14.520 toolkit. 05:15.240 --> 05:22.400 As you can see, it has been deprecated and you have to use the integrated Docker MCP toolkit in Docker 05:22.400 --> 05:25.800 Desktop 3.42 or later. 05:26.000 --> 05:31.480 So Docker MCP toolkit is a Docker desktop extension allowing allowing to connect Dockerized MCP servers 05:31.480 --> 05:32.560 to MCP clients. 05:32.840 --> 05:33.200 Yes. 05:33.200 --> 05:40.480 So a few days ago, the team integrated MCP toolkit into the user interface, and now it has a dedicated 05:40.520 --> 05:41.920 tab, as you can see here. 05:43.080 --> 05:47.320 Uh, it's still in beta, so you don't need to install this extension. 05:47.440 --> 05:50.480 You can go straight to the MCP toolkit. 05:51.120 --> 05:52.280 Now when you click on it. 05:54.840 --> 06:00.280 It gives you a whole set of tools to easily install and manage verified MCP servers. 06:01.560 --> 06:10.800 They have a catalog of nearly 120 MCP that are already set up, but I'm sure they will expand it soon. 06:11.560 --> 06:18.600 All of these run in isolated Docker containers, so you don't have to worry about them accessing anything 06:18.800 --> 06:20.680 on your system directly. 06:22.680 --> 06:25.120 So as the first step is dedicated for servers. 06:25.960 --> 06:27.600 The second one for MCP. 06:29.960 --> 06:31.480 And the last one for clients. 06:34.480 --> 06:43.920 So here you can connect all these MCP to apps like cloud desktop app, cursor or even some internal 06:43.920 --> 06:44.800 custom apps. 06:46.000 --> 06:46.480 All right. 06:46.480 --> 06:48.400 So now let's connect it to a client. 06:48.680 --> 06:52.800 So we are going to use Cloud Desktop app as our client as our MCP client. 06:53.480 --> 06:59.350 So you just simply need to click here to connect and don't need to edit configuration file like we did 06:59.390 --> 07:01.710 to connect the cloud desktop app to N810. 07:02.430 --> 07:07.670 And as I said, this works with other tools like cursor or even your own local setup. 07:08.270 --> 07:13.710 So you can browse and connect Dockerized MCP servers to your favorite MCP clients. 07:14.070 --> 07:20.910 So as you can see, you have access to some of the very popular apps like GitHub, like notion, like 07:21.310 --> 07:27.750 AWS, who can also give your agents access to MCP server like Azure. 07:28.230 --> 07:30.030 You also got 11 apps. 07:30.270 --> 07:36.270 You can use fire, which is really great to scrape the internet and Google Maps. 07:36.790 --> 07:40.710 And when you scroll down, I'm sure you will be familiar with some of the apps. 07:42.510 --> 07:45.670 We also have Shopify. 07:45.710 --> 07:46.310 Slack. 07:49.830 --> 07:50.670 Wolfram. 07:50.670 --> 07:51.470 Wikipedia. 07:52.190 --> 07:52.910 Webflow. 07:53.630 --> 07:57.190 Now when you click on the specific MCP server. 07:58.110 --> 08:00.350 Let's use Google Maps as an example. 08:04.310 --> 08:05.830 Here we have an overview. 08:06.550 --> 08:10.230 So this tool is for interacting with the Google Maps API. 08:10.870 --> 08:14.630 And here you have more details such as links to repository. 08:16.510 --> 08:17.830 And this is very important. 08:17.870 --> 08:18.870 Available tools. 08:19.430 --> 08:25.750 So here we can check what tools we can use for this specific MCP server. 08:26.030 --> 08:30.630 So in Google Maps we've got maps directions. 08:30.630 --> 08:36.670 To get directions between two points you have maps distance matrix to calculate travel distance and 08:36.670 --> 08:39.150 time for multiple origins and destinations. 08:40.430 --> 08:45.150 We also have maps, search places to search for, places using Google Places API. 08:46.950 --> 08:47.950 And when you scroll down. 08:50.430 --> 08:56.230 You have more details about the tools, which is really useful because by visiting this page. 08:56.270 --> 08:57.070 An overview. 08:57.710 --> 09:04.830 You can figure out what tools you would need to use for your specific use case. 09:05.510 --> 09:12.310 Next, when you go to tools tab, here you have a short descriptions of the tools the SMTP server offers. 09:13.830 --> 09:16.110 And then also very important configuration. 09:16.310 --> 09:20.350 So here in that case we just only need to provide API keys. 09:20.350 --> 09:25.550 Because to be able to use this server you have to provide API keys. 09:26.310 --> 09:33.550 And I'm going to show you how to find API keys for some of the services like Google Maps, perplexity, 09:33.590 --> 09:35.030 pickle and so on. 09:36.150 --> 09:40.470 If you want, you can go to repository to check the GitHub. 09:47.110 --> 09:52.830 And here you can read through the documentation to learn more about this particular Amqp server. 09:53.550 --> 09:58.950 So if you are interested, you can learn in more detail how it works and how to configure it manually. 09:59.910 --> 10:06.350 As you can see, this specific repo contains a collection of MCP servers. 10:07.350 --> 10:09.110 So for example, let's open Google Maps. 10:14.430 --> 10:16.270 And here we have some additional information. 10:17.230 --> 10:18.830 Google maps MCP server. 10:19.430 --> 10:22.630 For the Google Maps API we have tools. 10:22.670 --> 10:29.350 Again you have setup if you'd like to learn how to set up this manually. 10:32.470 --> 10:33.870 So you have JSON snippets. 10:34.590 --> 10:37.950 And on the left hand side we have all the files. 10:38.590 --> 10:42.630 So the code packages dependencies instructions and so on. 10:43.510 --> 10:43.790 Yes. 10:43.790 --> 10:52.150 So this MCP toolkit by Docker makes things so easy because the only thing you need to do is to for example 10:52.150 --> 10:53.110 go to stripe. 10:56.550 --> 10:59.310 Added to your MCP servers. 11:00.070 --> 11:03.590 But in order to do that, you have to provide credentials. 11:03.870 --> 11:05.870 So configure this MCP server. 11:06.070 --> 11:07.110 But this is really simple. 11:07.150 --> 11:09.150 You need to find API keys and that's it. 11:09.310 --> 11:11.550 So go to configuration here. 11:11.550 --> 11:12.790 Paste your API keys. 11:13.750 --> 11:15.190 And it's all set up for you. 11:15.190 --> 11:18.670 So you don't need to mess up with code and do everything manually. 11:19.350 --> 11:23.230 So this tool is really great because it can save you a lot of time and money. 11:23.710 --> 11:24.110 All right. 11:24.150 --> 11:27.430 Now let's use Cloud Desktop App as our MCP client. 11:27.790 --> 11:31.590 So once we install it go to Docker Desktop. 11:32.110 --> 11:34.830 Launch it go to settings. 11:44.150 --> 11:48.670 If you have never used Cloud Desktop app before you need to enable developer mode. 11:49.470 --> 11:54.230 There should be an option to switch it on, then go to developer. 11:56.590 --> 11:58.750 And here you should see all the integrations. 11:59.350 --> 12:01.390 For now I've got just an A10. 12:01.670 --> 12:04.030 So my cloud desktop app is connected to an A10. 12:05.150 --> 12:07.150 And now when you go to Docker Desktop. 12:13.950 --> 12:19.830 In the MCP toolkit, click on Clients Connect. 12:22.230 --> 12:25.710 Now your MCP toolkit should be connected to Cloud Desktop app. 12:26.270 --> 12:32.550 But before you will be able to see the connection inside Cloud Desktop app, you need to restart the 12:32.550 --> 12:33.950 Cloud desktop app completely. 12:34.750 --> 12:36.150 Just make sure to quit it. 12:36.830 --> 12:38.190 And this is really important. 12:38.230 --> 12:40.590 If you don't do that, it will not work. 12:41.670 --> 12:42.710 Now open it again. 12:43.590 --> 12:46.710 You can see that MCP Docker connection appeared. 12:47.510 --> 12:48.590 So let's open it. 12:49.830 --> 12:51.790 And just to confirm it's running now. 12:52.390 --> 12:58.670 So this means that our MCP toolkit has been successfully connected to cloud desktop app. 13:00.790 --> 13:09.630 Now, if for any reason your connection failed, just close the cloud desktop app, go to Docker Desktop. 13:19.190 --> 13:23.030 Press on this button again and then restart the cloud desktop app. 13:32.950 --> 13:39.910 And that's why Docker Desktop is so powerful, because you don't need to configure each MCP and each 13:39.910 --> 13:41.190 tool individually. 13:42.150 --> 13:51.270 You can simply connect all of them CP servers and the tools via Docker desktop and manage all of them 13:51.870 --> 13:52.710 From this up. 13:53.310 --> 13:53.910 All right. 13:53.950 --> 14:01.550 Now, once our MCP toolkit is connected to the cloud desktop app, let's start configuring MCP. 14:02.550 --> 14:04.630 So let's search for Google Maps. 14:09.630 --> 14:10.630 Configuration. 14:11.270 --> 14:19.230 And here we need to provide Google Maps API keys first before we start generating API keys and connecting 14:19.390 --> 14:21.550 applications to the MCP toolkit. 14:21.830 --> 14:25.550 Let's actually plan what we are going to do in the next part of this tutorial. 14:25.990 --> 14:29.070 So let's add a few MCP from a catalog. 14:33.790 --> 14:35.270 As our MCP servers. 14:38.750 --> 14:44.750 Because you can combine multiple MCP to complete a single task to get much better results. 14:45.470 --> 14:47.550 And that's why this is such a game changer. 14:49.190 --> 14:50.230 So this is our task. 14:50.230 --> 14:55.350 Let's say you are based in London and your remote team is mostly around Manchester. 14:56.150 --> 15:00.470 You are planning a weekend meet up somewhere close to Manchester where you can all work together in 15:00.510 --> 15:00.990 person. 15:01.830 --> 15:08.710 We'll use Google Maps to find towns within 35 minutes driving distance from Manchester. 15:10.750 --> 15:18.990 So when we search for Google Maps, MCP, the catalogue and go to tools. 15:21.590 --> 15:24.910 We'll use two tools offered by the MCP server. 15:26.230 --> 15:34.430 Let's use map search tool to find towns or areas nearby with coworking spaces or venues. 15:35.030 --> 15:42.590 And also let's use maps distance matrix to calculate actual driving times from Manchester, including 15:42.830 --> 15:43.910 traffic conditions. 15:45.070 --> 15:54.030 You can also use other tools, for example maps directions to get directions between two points or maps. 15:54.030 --> 15:58.390 Place details to get detailed information about a specific place. 15:58.630 --> 16:06.590 But for this, we are going to use other MCP perplexity to make this our server and use its tools. 16:07.870 --> 16:14.270 So then we can use perplexity MCP to research the pros and cons of each town, because we want to know 16:15.030 --> 16:22.510 if it's affordable, is it to get to, and if it's a good place for entertainment or to relax after 16:22.510 --> 16:22.870 work. 16:23.670 --> 16:30.550 Specifically, we are going to use Perplexity research tool to perform deep research using perplexity 16:30.630 --> 16:35.550 API and get a comprehensive research response. 16:37.070 --> 16:44.990 So yes, you can do all these things in one conversation inside the cloud desktop app without switching 16:45.070 --> 16:47.790 tabs, uh, or between websites. 16:48.270 --> 16:56.790 So you just install servers you need connect them to cloud and then I can start using them. 16:58.110 --> 17:01.430 So now I'm going to show you how to generate this API keys. 17:04.150 --> 17:07.070 First we need to create a project on Google Cloud. 17:08.190 --> 17:09.310 So search for console. 17:09.310 --> 17:10.110 Cloud Google. 17:11.590 --> 17:12.750 Open the first link. 17:16.550 --> 17:18.070 Then login to your Google account. 17:18.830 --> 17:19.950 Go to your console. 17:24.110 --> 17:25.950 Now here you can create a new project. 17:30.190 --> 17:31.430 So press on. 17:31.430 --> 17:32.510 Create new project. 17:35.030 --> 17:35.830 Give it a name. 17:46.430 --> 17:47.150 Create. 17:52.550 --> 17:57.030 Now just wait a moment, then select the project you just created. 18:00.790 --> 18:02.710 And make sure you are in the right project. 18:04.070 --> 18:09.590 Then what we need to do we need to enable APIs we want to use. 18:10.470 --> 18:13.230 So in that case we want to use. 18:14.830 --> 18:15.670 Let's check. 18:18.270 --> 18:22.910 Two tools maps search places and maps distance matrix. 18:23.390 --> 18:24.590 So let's search for it. 18:27.390 --> 18:30.630 So I'm going to type maps search. 18:31.710 --> 18:33.150 And this is the first result. 18:33.150 --> 18:34.270 So places API. 18:35.110 --> 18:36.110 So click on that. 18:41.630 --> 18:43.310 Then we need to enable API. 18:45.470 --> 18:46.390 So click enable. 18:50.590 --> 18:54.910 So you'll be able to get detailed information about 100 million places. 19:02.070 --> 19:03.510 Now let's copy the API key. 19:05.190 --> 19:06.790 Let's go to Google Maps platform. 19:09.550 --> 19:11.910 Here you can also restrict key. 19:12.670 --> 19:14.310 I'm going to select maybe later. 19:16.870 --> 19:19.030 Now when I go to APIs and services. 19:22.190 --> 19:28.630 I can see all the tools I can use from Google Maps platform. 19:30.310 --> 19:31.430 And when I scroll down. 19:37.790 --> 19:42.630 I can see that places API is enabled because I can disable it right here. 19:44.830 --> 19:52.580 Now we also want to Activate the Distance Matrix API to click on enable it. 19:57.100 --> 20:00.580 If you want to calculate the travel time and distance for multiple destinations. 20:02.340 --> 20:03.220 Alright, perfect. 20:04.340 --> 20:06.820 Now let's go back to the Docker desktop. 20:10.660 --> 20:16.500 And under configuration I'm going to remove this API key and creating one. 20:19.580 --> 20:22.380 Let's paste the API keys we copied. 20:25.180 --> 20:25.900 Confirm. 20:27.620 --> 20:28.020 All right. 20:28.020 --> 20:28.580 Perfect. 20:29.580 --> 20:35.580 And now we should have access to this MCP in our cloud desktop app. 20:36.580 --> 20:38.100 But we need to restart the app. 20:39.300 --> 20:40.100 So let's do it. 20:48.900 --> 20:52.500 And now every large language model offered by anthropic. 20:54.700 --> 20:56.820 Should have access to this MCP. 20:57.620 --> 21:01.020 So here you can switch between different large language models. 21:01.020 --> 21:06.020 You have access to cloud sonnet for cloud opus for animal models. 21:07.820 --> 21:13.460 And now when you press on this button you should see that you have access to MCP Docker. 21:14.140 --> 21:15.460 And when you expand it. 21:19.060 --> 21:23.740 These are all the tools that I connected in Docker desktop. 21:24.540 --> 21:29.540 So all MCP servers with their tools are connected to cloud desktop app. 21:29.900 --> 21:35.260 So here I have a file called Google Maps and Perplexity. 21:35.780 --> 21:40.260 And I can easily switch on and off each tool separately. 21:41.300 --> 21:49.780 So here we have our maps distance matrix And map search places where I. 21:49.820 --> 21:50.380 Perfect. 21:51.700 --> 21:53.420 Now let's actually give it a test. 21:53.860 --> 21:57.060 So I'm going to open the cloud desktop app again. 22:02.020 --> 22:08.260 And I'm going to ask it to use Google Maps tool to find out what cities near Manchester can I reach 22:08.260 --> 22:12.540 within 35 minutes driving during evening rush hour traffic? 22:14.300 --> 22:15.140 Let's send it. 22:20.660 --> 22:21.740 And let's see what happens. 22:23.260 --> 22:24.580 So I will help you find. 22:24.620 --> 22:32.260 Okay, so now cloud is asking about permission to use this MCP server. 22:33.020 --> 22:36.660 So as you can see maps search places MCP Docker. 22:37.820 --> 22:39.060 So I'm going to allow once. 22:42.900 --> 22:46.700 Let me search for nearby places and then check driving times. 22:46.740 --> 22:47.180 Okay. 22:48.260 --> 22:51.540 Now it's going to use this MCP server again. 22:51.900 --> 22:53.860 So this time I'm going to allow always. 22:53.860 --> 22:55.980 So you don't need to give it permission every time. 23:01.580 --> 23:06.580 Now cloud is asking for the permission to use our second tool. 23:06.740 --> 23:11.340 So maps distance matrix from this MCP server. 23:13.460 --> 23:15.340 Okay I'm going to allow always. 23:18.420 --> 23:18.660 Okay. 23:18.660 --> 23:19.900 Now let's wait a moment. 23:27.140 --> 23:27.660 Perfect. 23:27.700 --> 23:28.940 Now we got results. 23:30.020 --> 23:34.420 Here you can see the entire process of performing this search. 23:34.940 --> 23:41.740 So I will help you find cities in Manchester that are reachable within 35 minutes during rush hour. 23:41.780 --> 23:46.740 Let me search for nearby places and then check driving times during typical rush hour conditions. 23:47.220 --> 23:54.980 And here you can see that cloud was using Maps Places tool from the Google Maps. 23:55.020 --> 24:00.060 And you can also expand it to see the request. 24:01.780 --> 24:02.940 You can see the response. 24:06.460 --> 24:06.820 Right. 24:06.820 --> 24:07.340 Great. 24:08.540 --> 24:13.060 Then let me search for more specific cities and towns around Manchester. 24:14.060 --> 24:18.980 It was using map search places tool three times. 24:23.580 --> 24:24.580 Okay, great. 24:24.580 --> 24:32.620 And then now let me check driving times from Manchester to these nearby cities during evening rush hour. 24:32.820 --> 24:36.180 I will use the distance matrix to get multiple destinations at once. 24:36.820 --> 24:38.940 And here it was using the second tool. 24:38.940 --> 24:40.420 So maps distance matrix. 24:40.940 --> 24:46.500 Let's let's expand it more Of driving college in Manchester, UK. 24:46.620 --> 24:47.660 Destinations. 24:49.740 --> 24:57.740 Okay, so it's found quite a few cities based on my search using Google Maps data. 24:57.860 --> 25:03.460 Here are the cities and towns in Manchester that you can reach within 35 minutes of driving during typical 25:03.460 --> 25:04.540 traffic conditions. 25:05.300 --> 25:07.420 And here you have the list of the cities. 25:09.340 --> 25:10.020 All right. 25:10.180 --> 25:12.300 Just outside the 35 minute range. 25:12.900 --> 25:18.220 Please note that these times are based on current traffic data and may vary during actual evening rush 25:18.220 --> 25:18.420 hour. 25:18.460 --> 25:23.500 Typically 5 to 7 p.m. during heavy rush hour traffic. 25:23.500 --> 25:29.260 Some of these journeys could take longer, particularly those involving major motorways like right. 25:29.260 --> 25:34.140 For the most accurate real time estimates during your specific travel time, I would recommend checking 25:34.140 --> 25:36.740 Google Maps or similar navigation apps just before your journey. 25:37.380 --> 25:45.980 As you can see, cloud was using current real time traffic data and the output is pretty detailed and 25:46.220 --> 25:49.660 accurate in the second part of this task. 25:51.060 --> 25:53.980 Now let's ask perplexity. 25:54.660 --> 25:55.860 So use perplexity. 26:00.420 --> 26:01.500 Let me just check. 26:03.780 --> 26:04.580 Okay. 26:06.900 --> 26:09.460 To research the pros and cons of each town. 26:12.540 --> 26:12.980 Okay. 26:13.020 --> 26:14.140 Let's just copy it. 26:18.460 --> 26:19.460 And paste it here. 26:24.460 --> 26:25.260 Now let's see. 26:26.380 --> 26:29.100 I will use perplexity to research the pros and cons of each town. 26:30.300 --> 26:32.180 Now it's gonna ask for permission. 26:32.540 --> 26:33.260 Going to error. 26:33.300 --> 26:33.900 Always. 26:37.980 --> 26:43.660 I think it will take maximum two minutes, because perplexity is going to do a deep research. 26:43.660 --> 26:43.660 Deepresearch. 26:49.020 --> 26:50.780 Let's give it some time to process it. 26:54.180 --> 26:58.580 All right, now let's check the response based on my research using perplexity. 26:58.620 --> 27:02.460 Here are comprehensive analysis of the pros and cons of each town in Manchester. 27:03.060 --> 27:05.300 Salford 70 minutes from Manchester. 27:05.300 --> 27:12.780 Pros excellent transport links with frequent trams every ten minutes and direct Metrolink connections. 27:12.820 --> 27:23.660 Cons most expensive for solo living utilities limited local options okay residence okay Stockport. 27:23.700 --> 27:24.780 Good transport. 27:27.060 --> 27:27.780 All right. 27:30.100 --> 27:30.660 Outcome. 27:30.700 --> 27:31.620 Very affordable. 27:38.300 --> 27:39.700 Works and transport. 27:40.500 --> 27:40.900 Limited. 27:40.900 --> 27:42.300 Evening entertainment locally. 27:49.020 --> 27:51.380 Chi Minh infrastructure. 27:52.580 --> 27:54.020 Best local entertainment. 27:56.860 --> 27:57.220 All right. 27:57.220 --> 27:57.660 Summary. 27:57.700 --> 28:00.780 Recommendations for affordability. 28:00.820 --> 28:04.500 Bolton or Oldham offer the best value for quick commute. 28:06.020 --> 28:15.580 33 minutes or 30 minutes, or fastest for entertainment or income has the best local scene Bolton for 28:15.580 --> 28:15.980 students. 28:16.020 --> 28:22.660 Nightlife nightlife best overall balance value offers good transport, moderate costs and reasonable 28:22.900 --> 28:23.660 amenities. 28:24.100 --> 28:30.260 Avoid budget conscious Salford and income are significantly more expensive. 28:30.820 --> 28:37.500 The choice depends on your priorities, whether you value short commutes, local entertainment or keeping 28:37.500 --> 28:38.180 costs low. 28:39.180 --> 28:39.740 Perfect. 28:39.740 --> 28:46.020 So we are able to get real time data from Google Maps using the Google Maps MCP. 28:48.340 --> 28:52.660 The actual driving times, not just rough estimates. 28:53.180 --> 29:00.220 Then we used one of the best AI search engines on the market to perplexity, to do deep research on 29:00.260 --> 29:06.220 each location and all of that just by asking a single question inside the cloud app. 29:10.460 --> 29:12.540 Now, this was just an example. 29:12.580 --> 29:19.260 Now imagine how much time you could save when you start combining even more MCP servers and tools, 29:19.460 --> 29:23.820 especially as Docker team the continuous verifying more of them. 29:24.580 --> 29:31.660 All right, now I'm going to show you how to generate API keys in perplexity so you can set up perplexity 29:31.660 --> 29:32.900 as your MCP server. 29:33.420 --> 29:35.740 So first go to perplexity comm. 29:36.100 --> 29:42.220 If you haven't used perplexity before, just for the context, it's one of the best AI powered search 29:42.220 --> 29:42.820 engines. 29:43.020 --> 29:49.500 Right now on the market, it takes your question, searches the web in real time, and then gives you 29:49.500 --> 29:52.020 a clear, concise answer. 29:52.460 --> 29:55.580 Usually the answer is backed by resources. 29:55.820 --> 29:57.540 You can click on verify. 29:58.260 --> 30:01.460 Also, it doesn't hallucinate as much as other llms. 30:01.740 --> 30:08.020 So the perplexity AI is an advanced AI powered web search engine designed to provide accurate, trusted 30:08.020 --> 30:14.300 and real time answers to user to user questions by combining large language models with live internet 30:14.300 --> 30:14.780 search. 30:15.260 --> 30:21.500 It uses sophisticated natural language processing to understand queries contextually, and then searches 30:21.500 --> 30:28.340 authoritative online sources to synthesize concise, human like responses, complete with citations 30:28.340 --> 30:30.380 to the original sources for verification. 30:31.180 --> 30:38.580 Founded in August 2022, perplexity AI has quickly grown into a significant player in the AI search 30:38.580 --> 30:46.260 space valued at around $14 billion as of mid 2025, with a user base processing tens of millions of 30:46.260 --> 30:47.100 queries daily. 30:47.460 --> 30:53.060 The company offers a free version and paid subscriptions, which unlock more advanced AI models and 30:53.060 --> 30:58.700 features such as multimodal inputs, image documents, follow up questions, capabilities and enhanced 30:58.700 --> 30:59.780 search depth. 31:00.100 --> 31:07.140 And the key features conversational search so allows users to ask follow up questions and maintain context 31:07.180 --> 31:10.980 within a session, making it more interactive than traditional search engines. 31:11.460 --> 31:12.620 Return data access. 31:12.620 --> 31:19.380 Unlike many AI models limited to static training data, perplexity assesses live web data enabling app 31:20.740 --> 31:28.020 enabling up to date answers including live sports scores, election results and news events to search 31:28.060 --> 31:28.780 modes. 31:29.100 --> 31:35.740 Quick search for fast, straightforward answers, plus search for deeper, more tailored responses that 31:35.740 --> 31:37.740 may involve clarifying questions. 31:38.140 --> 31:39.340 All right, how it works. 31:39.980 --> 31:43.980 It uses a blend of proprietary and third party llms, including GPT four. 31:46.540 --> 31:53.900 GPT three and its own sonar model to interpret queries, retrieve relevant information from the web, 31:53.900 --> 31:56.420 and generate coherent summaries. 31:56.660 --> 32:01.380 It continuously crawls and indexes web content to keep its knowledge base current. 32:02.380 --> 32:08.140 In use cases, perplexity AI is used for research, content summarization, language translation, data 32:08.180 --> 32:12.220 analysis, and general information retrieval across various industries. 32:12.260 --> 32:18.540 It is especially valued for providing verifiable, up to date information quickly and efficiently. 32:18.860 --> 32:25.100 In summary, perplexity is a next generation AI search engine that combines conversational AI with real 32:25.140 --> 32:31.100 time web search to deliver accurate source answers, helping users explore and verify information more 32:31.100 --> 32:35.100 efficiently than traditional search engines or standalone AI chatbots. 32:35.540 --> 32:36.860 Perplexity is very powerful. 32:36.900 --> 32:43.970 I personally prefer to use it to ask question to search the web instead of going to ChatGPT or cloth 32:43.970 --> 32:47.330 and ask the models to get an answer? 32:48.170 --> 32:48.530 Yes. 32:48.530 --> 32:54.730 So I usually go to perplexity first when I have a question, because it pulls real time info from the 32:54.730 --> 33:02.010 web so you can get more current and accurate responses than you would from ChatGPT or cloth. 33:02.050 --> 33:07.370 Even if you are using the search web features also without the extra fluff. 33:07.850 --> 33:16.570 So in order to connect perplexity to MCP toolkit in Docker Desktop, let's go to Perplexity Comm. 33:17.530 --> 33:18.610 Create your account. 33:18.770 --> 33:19.850 I'm going to sign in. 33:22.450 --> 33:24.850 Then go to your account at the bottom left corner. 33:27.570 --> 33:28.490 Click on account. 33:32.050 --> 33:32.970 API keys. 33:37.330 --> 33:42.010 But before you will be able to generate your API keys, you have to top up your account. 33:44.330 --> 33:45.330 Then create key. 33:49.170 --> 33:50.010 Copy it. 33:50.930 --> 33:53.770 Go back to Docker Desktop. 33:59.610 --> 34:00.570 Go to catalog. 34:02.170 --> 34:03.370 Search for perplexity. 34:06.450 --> 34:07.690 Go to configuration. 34:09.330 --> 34:13.530 I'm going to delete the API keys I already have and create new one. 34:14.770 --> 34:19.490 Paste your perplexity API key here and confirm. 34:20.930 --> 34:21.290 Great! 34:21.290 --> 34:24.850 And now when you go to Cloud Desktop app. 34:27.250 --> 34:28.370 And go to tools. 34:31.010 --> 34:36.610 In the MCP Docker you should be able to see three tools from perplexity. 34:36.650 --> 34:39.730 So perplexity ask perplexity, reason and perplexity research. 34:40.210 --> 34:42.130 So these are exactly the same tools. 34:45.130 --> 34:46.210 You can see here. 34:48.970 --> 34:52.530 So now you are able to use perplexity inside Cloud desktop app. 34:53.970 --> 34:59.610 Now just to let you know, using perplexity research is quite expensive because it uses deep research 35:00.130 --> 35:04.170 to find relevant information for for your questions. 35:04.890 --> 35:06.010 So just keep it in mind. 35:07.170 --> 35:11.570 Now I'm going to show you how to connect another popular tool if I recall. 35:12.130 --> 35:13.970 So it's kind of like perplexity. 35:13.970 --> 35:21.970 But instead of giving you answers, it scrapes retail websites and pulls and pulls out raw data. 35:22.610 --> 35:29.730 So perplexity gives you ready answers based on data, while five crore gives you data to work with. 35:30.650 --> 35:32.530 So let's go to File Crawl. 35:34.570 --> 35:40.490 As you can see, this is the fast, reliable web scraper for LMS, you can power your AI apps with clean 35:40.490 --> 35:42.130 data cloud from any website. 35:42.810 --> 35:45.530 It's also open source, the same as N810. 35:46.530 --> 35:49.370 You can start for free with 500 credits. 35:51.970 --> 35:53.650 It's trusted by the top companies. 36:06.170 --> 36:08.650 As you can see, you can add it to m.c.p.s. 36:16.850 --> 36:18.530 All right, so let's create an account. 36:22.050 --> 36:22.690 Sign up. 36:25.770 --> 36:28.050 Provide email and password and then login. 36:28.690 --> 36:30.850 I'm going to login because I already have an account. 36:31.970 --> 36:33.930 Then simply go to API keys. 36:36.290 --> 36:37.810 Here you can create your API key. 36:39.570 --> 36:47.410 So give it a name for example Docker desktop and create. 36:51.250 --> 36:51.890 Copy. 36:53.450 --> 36:54.810 Go to Docker desktop. 36:56.850 --> 36:58.930 Search for file crawl. 37:05.850 --> 37:07.170 Go to configuration tab. 37:08.810 --> 37:09.810 Go to secrets. 37:14.690 --> 37:25.610 I'm going to remove the current API key and create new one just to demonstrate to you and confirm save. 37:26.330 --> 37:27.130 And that's it. 37:27.490 --> 37:31.330 It should be set up and you should be able to use it in your Google Cloud desktop. 37:32.530 --> 37:37.570 Actually let's check and you should have access to it. 37:37.610 --> 37:38.050 Tools. 37:40.530 --> 37:41.890 You can read the descriptions. 37:44.250 --> 37:45.770 Go to your desktop app. 37:45.930 --> 37:48.090 Let's check tools. 37:50.290 --> 37:50.730 Mocp. 37:50.770 --> 37:51.250 Docker. 37:51.770 --> 37:55.410 And as you can see now I can use all of the tools. 37:58.130 --> 37:58.490 All right. 37:58.490 --> 37:59.090 Perfect. 38:00.130 --> 38:00.610 All right. 38:00.610 --> 38:00.890 That's it. 38:00.890 --> 38:08.330 For this tutorial you learned how to use how to combine tools like Google Maps, perplexity and through 38:08.370 --> 38:15.290 MCP toolkit in in Docker and how how you can use it in cloud desktop app and how easy it is to build 38:15.290 --> 38:16.290 something useful. 38:16.850 --> 38:23.490 I strongly recommend you using MCP Toolkit in Docker Desktop because it can save you a lot of time, 38:23.890 --> 38:32.610 and also it can open a lot of possibilities, especially as more verified MCP get added over time. 38:33.370 --> 38:34.610 I hope this was helpful. 38:34.850 --> 38:36.930 Thanks for watching and see you in the next one.