WEBVTT 00:00.200 --> 00:02.400 But as I say, it's quite interesting that that happened. 00:02.440 --> 00:07.920 It gave us an opportunity to understand more about remote servers and their stability and the like. 00:07.920 --> 00:08.280 Okay. 00:08.320 --> 00:12.160 So now the final one I want to show you is completely optional. 00:12.400 --> 00:15.600 I have done this one for myself and I've enjoyed it. 00:15.640 --> 00:18.160 You do not need to for one moment. 00:18.360 --> 00:21.200 So please take this as an optional thing. 00:21.720 --> 00:28.920 There is a service called, uh, that's that's, uh, it's called Financial Datasets AI. 00:29.400 --> 00:33.760 And it is a provider of real time financial information. 00:34.000 --> 00:35.840 But that does come at a price. 00:36.040 --> 00:39.600 So they charge about $0.01 for every lookup. 00:39.800 --> 00:42.360 And they are they're a good company. 00:42.360 --> 00:48.920 They are a, uh they're also part of the recommended MCP servers from, from anthropic. 00:48.920 --> 00:54.080 So they direct you to this and it's used by, uh, lots of data scientists. 00:54.080 --> 00:59.240 So it's active and the data science community, uh, and, um, they are they're yeah. 00:59.280 --> 01:05.200 As I say, they're perfectly, uh, decent, but they are a little bit pricey because $0.01 per lookup 01:05.320 --> 01:09.000 doesn't sound like it's going to break the bank, but if you're going to do a lot of looking up, then 01:09.000 --> 01:09.520 it could do. 01:09.520 --> 01:14.080 And also they have a minimum $20 that you have to put there to then draw down against. 01:14.080 --> 01:15.880 So I put in my $20. 01:15.920 --> 01:19.960 I haven't yet got very far through it, but I've been enjoying having that $20. 01:19.960 --> 01:22.440 You should only do this if you're comfortable with that. 01:22.440 --> 01:26.720 And we're going to be building something that we'll be looking and trying to simulate trading financial 01:26.720 --> 01:27.080 markets. 01:27.080 --> 01:31.120 And for me, it was well worth 20 bucks to be able to to play with this. 01:31.120 --> 01:32.680 But it's a preference thing. 01:32.960 --> 01:35.440 You can stick with a free service which is down right now. 01:35.440 --> 01:40.080 When it's running, you will get free stock prices, but you'll only get I think it's 25 a day. 01:40.200 --> 01:41.280 It's something like that. 01:41.480 --> 01:45.320 Uh, and then you'd have to make another key if you wanted to, which you can keep doing, but it's 01:45.320 --> 01:46.160 a bit of a bore. 01:46.520 --> 01:49.240 Uh, so, uh, yeah, there's the free option. 01:49.280 --> 01:51.360 Should you should you be okay to do that? 01:51.360 --> 01:57.000 But if you don't mind forking out 20 bucks or your local equivalent, then this is a great service. 01:57.000 --> 02:03.840 And by the way, you should go and explore on Macpaw or one of the others, uh, glamour AI and see 02:03.840 --> 02:09.360 if you can find a better provider of financial data that you prefer at a price point that you like, 02:09.450 --> 02:11.570 but I've gone ahead with the $20. 02:11.810 --> 02:17.410 Now, the other thing about this one is that there's a slightly more tricky process for setting it up, 02:17.410 --> 02:22.250 but that is good because it gives us a chance to experiment with with another way that this is sometimes 02:22.250 --> 02:22.690 done. 02:22.810 --> 02:24.450 So let me take you through that. 02:24.450 --> 02:29.410 And even if you decide you don't want to pay the 20 bucks, it's still worth you seeing this process 02:29.410 --> 02:32.490 because it's a it is a fairly common kind as well. 02:33.090 --> 02:39.290 In the past, we've used commands like UVF and Mpcs which allow you to run. 02:39.330 --> 02:46.930 UVF lets you run like a tool, um, a package that's, uh, that's Pip installed, but it's not actually 02:46.930 --> 02:47.690 installed. 02:47.730 --> 02:50.330 It's able to run it without needing to install it. 02:50.330 --> 02:52.970 But this, this tool works differently. 02:52.970 --> 02:58.610 The financial data service, they need you if we if we go to their their entry on MCP. 02:58.930 --> 03:00.930 So let me show it to you. 03:01.170 --> 03:07.530 So this is the financial data sets MCP server that connects to their service and what they need you 03:07.530 --> 03:07.810 to do. 03:07.850 --> 03:09.650 These are these are the tools that they provide. 03:09.650 --> 03:14.410 Get income statements, get balance sheets, cash flow current price prices and news. 03:14.410 --> 03:17.370 So they've got a lot of stuff here that's like really great. 03:17.370 --> 03:18.730 And it's real time. 03:19.050 --> 03:26.890 Um, so you have to, uh, install it properly as a project as a, as a Python project. 03:26.890 --> 03:35.690 So first you'll go in and type git clone and the, the, the, um, the folder, uh, and do this in 03:35.690 --> 03:36.970 your projects folder. 03:36.970 --> 03:40.050 So in your home directory you have a projects uh folder. 03:40.050 --> 03:42.170 And you will clone this in there. 03:42.450 --> 03:45.370 Um, we, it says if you don't have you've installed install it. 03:45.370 --> 03:47.930 But luckily we do have, you've installed of course. 03:48.210 --> 03:51.930 And then you just follow their instructions, do exactly what they say. 03:52.250 --> 04:02.290 Uh, then they need you to create a local file in their, in their project in this MCP server project. 04:02.530 --> 04:08.530 And you populate that with the financial data sets API key that we are about to set up. 04:08.850 --> 04:11.210 And then that's all you don't need to do that last step. 04:11.210 --> 04:13.210 That is all that you need to do. 04:13.250 --> 04:16.850 And by the way, you'll be able to give Claude access to this as well, should you wish. 04:16.850 --> 04:20.420 Claude desktop so you could be talking with Claude about the stock market. 04:20.420 --> 04:21.460 That's like an added boost. 04:21.500 --> 04:25.180 Of course, you could do that to all of the MCP server tools that we've been talking about. 04:25.540 --> 04:26.140 All right. 04:26.140 --> 04:32.740 So if I come back here, if I just bring up the terminal, uh, I just want to show you that this is 04:32.740 --> 04:38.940 if I go up a directory and look in my projects directory, you will see that there is indeed an MCP 04:38.980 --> 04:42.540 server directory there because I cloned in there. 04:42.540 --> 04:43.780 And this is it. 04:43.820 --> 04:49.780 And there is if I look at hidden files as well, you'll see that there is a dot env file that I've created 04:49.780 --> 04:51.180 and that I've populated. 04:51.180 --> 04:58.940 So this is now set up and running for me, installed locally as another project alongside the agents 04:58.940 --> 04:59.660 project. 04:59.660 --> 05:05.180 And so if you do wish to splash out and be part of this real time market data, then the next step is 05:05.180 --> 05:10.820 to go to financial datasets where you can sign up for an account, which I've done. 05:10.980 --> 05:16.060 And when you come in, you get a dashboard after you've signed up like this and you can see I made a 05:16.060 --> 05:18.860 bunch of requests as I've been trying different things. 05:19.100 --> 05:23.060 And when you go to billing here and to credits. 05:23.060 --> 05:27.300 This is where you manage putting a bit of a top up on your account. 05:27.300 --> 05:31.220 It takes a $20 minimum and you can see here the costs of API's. 05:31.260 --> 05:34.860 It's free to get company facts and earnings releases. 05:34.900 --> 05:37.580 A stock price is $0.01 per request. 05:37.620 --> 05:40.460 Crypto prices is $0.01 per request. 05:40.460 --> 05:42.820 And then there's some other things down here as well. 05:42.820 --> 05:46.060 But some of them I don't think are exposed through the MCP server. 05:46.580 --> 05:52.340 So that is that is what you can can look to set up and create an API key should you wish to. 05:52.620 --> 05:59.500 And then once you've done that, then as usual, you put that API key into the EMV file. 05:59.500 --> 06:00.860 In Financial Data sets API key. 06:00.900 --> 06:03.540 You actually have to put that into EMV files. 06:03.540 --> 06:04.900 In are EMV file. 06:04.900 --> 06:12.140 That's right here, because we'll use it directly in our accounts code, but also in the env file of 06:12.140 --> 06:15.180 the project where the MCP server is running. 06:15.220 --> 06:15.580 Okay. 06:15.620 --> 06:17.860 And with this it is time to run this. 06:17.940 --> 06:23.620 So the way that we actually specify the MCP tool to run is a bit different now. 06:23.620 --> 06:25.740 We're still using the second technique here. 06:25.740 --> 06:32.230 So we're still running a local MCP server that's going to call remotely, but we're running it based 06:32.230 --> 06:35.790 on a project that we have downloaded using git clone. 06:35.790 --> 06:40.830 So the way we have to do this is we specify the directory where this project is. 06:40.870 --> 06:46.230 It's right here for me in AD projects MCP server for you it will be wherever you've put it. 06:46.230 --> 06:50.390 And if you're on a PC there'll be backslashes and uh, but it should probably be. 06:50.430 --> 06:53.110 It will be called MCP server if you cloned it that way. 06:53.590 --> 06:56.870 We need to pass in the key like this. 06:56.870 --> 07:01.710 Now, it's interesting because we did set a env file in this project, but it turns out that you need 07:01.710 --> 07:03.150 to do this for it to work. 07:03.150 --> 07:05.750 So it may be that it doesn't use the env file, but. 07:05.750 --> 07:07.150 But I followed its instructions. 07:07.150 --> 07:08.150 I put it there too. 07:08.390 --> 07:10.030 So we'll specify it this way. 07:10.270 --> 07:15.070 And then we, we specify uh, the parameters to run this. 07:15.070 --> 07:20.070 And the way we run it, it isn't us this time because we're just running their Python code. 07:20.070 --> 07:26.670 So it's UV minus minus directory and then the directory where it is run server.py. 07:26.830 --> 07:31.030 And if you look in this directory you'll see that there is a Python script Server.py. 07:31.150 --> 07:35.390 That launches their server and we'll pass in our environment variable. 07:35.390 --> 07:36.510 So we will run this. 07:36.830 --> 07:41.470 And then the code the MCP server code is identical to before. 07:41.470 --> 07:42.350 So simple. 07:42.470 --> 07:44.390 Let's see what tools we get access to. 07:44.430 --> 07:50.150 We get tools to get income statements, balance sheets, cash flow, historic prices, and a bunch of 07:50.150 --> 07:52.550 crypto stuff too, if you're interested. 07:52.830 --> 08:01.510 And then we will quite simply call our MCP server the same way as always, we will pass in our requests 08:01.550 --> 08:03.390 to get Amazon's share price. 08:03.390 --> 08:08.590 This time, we're going to be spending a cent for the privilege of learning Amazon's share price. 08:08.590 --> 08:10.230 But it does come through. 08:10.390 --> 08:11.790 We get the share price. 08:11.790 --> 08:16.950 It's $170 apparently, and it's dropped today 2.45% drop. 08:17.230 --> 08:20.230 So we get a bunch of extra information to boot. 08:20.510 --> 08:26.710 And that concludes our last of our MCP servers for today. 08:27.150 --> 08:30.950 But the exercises for you, of course, is to explore more. 08:31.150 --> 08:35.200 You now know all of the techniques for hooking up to MCP servers. 08:35.200 --> 08:39.320 Which means that, yeah, the world is open to you. 08:39.560 --> 08:40.440 Do some browsing. 08:40.480 --> 08:46.280 Look on the marketplace, find some interesting looking MCP servers, and then put them to good use. 08:46.600 --> 08:51.320 Uh, build an agent, ask them some questions and explore and enjoy it. 08:51.360 --> 08:56.560 And put some examples in community contributions, or post them on on LinkedIn so I can weigh in. 08:56.560 --> 08:58.360 I would love to see what you're doing. 08:58.520 --> 09:06.320 The the joy of this is that it opens up so many capabilities that you can easily integrate with your 09:06.320 --> 09:07.120 own agent. 09:07.120 --> 09:08.000 So go do it. 09:08.000 --> 09:09.200 And so there you have it. 09:09.240 --> 09:11.600 We've done one, two and three. 09:11.840 --> 09:18.360 And for number two we did a couple of different ways of, of uh, of, of using, of calling the code 09:18.360 --> 09:22.080 one where we did a git clone and ran it the whole thing locally. 09:22.720 --> 09:27.200 And that wraps up day three of week six. 09:27.240 --> 09:28.760 We've only got two days left. 09:28.760 --> 09:31.000 We're coming onto the very, very home stretch. 09:31.240 --> 09:37.200 And I want to make sure that we that we bring it home with a brilliant, brilliant final project. 09:37.240 --> 09:38.240 Great fun. 09:38.360 --> 09:40.120 I can't wait to tell you about it next time.