WEBVTT

00:00.440 --> 00:01.880
Welcome back to the course.

00:01.920 --> 00:07.440
Today I'm going to show you how to set up and use AP files your MCP server.

00:07.720 --> 00:12.960
And honestly, I think this is going to completely change how you approach data scraping.

00:14.160 --> 00:20.520
Basically, you will be able to write a single prompt and have AI pull live data from practically anywhere,

00:21.200 --> 00:25.320
including Instagram, LinkedIn, Google Maps, you name it.

00:26.280 --> 00:30.920
Just tell it exactly what you want to scrape and it handles everything for you.

00:31.520 --> 00:35.280
Now, the best part about this setup is there is no coding involved.

00:35.760 --> 00:41.320
So you give it a simple instruction and it goes out and gets the data in exactly the format you need.

00:41.600 --> 00:45.000
So we are talking about getting this running just in just a few minutes.

00:46.480 --> 00:50.320
You just ask AI to do it and it knows its tools to use.

00:53.200 --> 00:57.120
How to structure the request and delivers organized data back to you.

00:59.160 --> 01:04.560
What makes this powerful is that it can combine multiple data sources in a single request.

01:04.560 --> 01:11.400
So you could ask it to scrape Instagram posts about a specific topic, then find related LinkedIn profiles

01:11.960 --> 01:15.520
and cross-reference that with Google Maps business data all in one go.

01:17.600 --> 01:24.240
Now, the traditional way of doing this would require multiple tools and separate API keys for each

01:24.240 --> 01:27.560
of these services, and a lot of manual setup.

01:28.160 --> 01:33.120
With this setup, it's all handled automatically through one interface.

01:34.280 --> 01:40.080
Let me show you exactly how this works, and we will start with the basics of what Apify is and why

01:40.440 --> 01:42.640
this integration is such a game changer.

01:44.000 --> 01:47.440
Each scraper has its own pricing structure.

01:47.440 --> 01:49.520
For example, this scraper.

01:52.960 --> 01:57.040
Cost $6 for every thousand results you extract.

01:58.720 --> 02:04.280
And the good news is you can test these scrapers directly on the website for free.

02:05.840 --> 02:14.520
For example, if you take this URL and click start, it's going to run the operation and start pulling

02:14.520 --> 02:16.680
product data from Amazon's website.

02:18.200 --> 02:20.080
As you can see, the actor is getting your data.

02:21.160 --> 02:24.040
It usually takes around one minute to start seeing results.

02:25.320 --> 02:30.200
So you can see it's working through different products right now and we are getting structured data

02:30.240 --> 02:30.640
back.

02:31.720 --> 02:37.920
We got product titles, brand names and the actual URLs for the for each product.

02:38.600 --> 02:42.040
So this is exactly the kind of data you will get from most scrapers.

02:43.240 --> 02:49.520
Clean, organized information that you can feed into whatever language model you are using for analysis,

02:49.680 --> 02:51.720
or any other purpose you have in mind.

02:53.640 --> 03:00.200
But for example, if you wanted to use Google Maps to extract data from thousands of locations and businesses.

03:00.800 --> 03:07.600
Things like reviews, reviews, reviewer details, images, contact info, opening hours, location,

03:07.600 --> 03:09.360
data prices and so on.

03:10.320 --> 03:16.280
You would need to pay per event, which means you are charged for each business location you scrape

03:16.280 --> 03:16.920
data from.

03:18.920 --> 03:24.720
So when you click on pricing, you will see it's paid per event so you are not charged for the Apify

03:24.880 --> 03:28.200
platform usage, just a fixed price for specific events.

03:28.920 --> 03:33.600
So as you can see, the base cost is 0.004 per place scraped.

03:36.840 --> 03:45.560
And there are additional costs for extras like 0.001 for filter applied, 0.002 for additional price

03:45.560 --> 03:46.160
details.

03:46.560 --> 03:49.040
So you pay for exactly what you extract.

03:50.920 --> 03:55.400
So the pricing varies depending on what you are extracting and how much detail you need.

03:55.440 --> 04:01.290
But the key thing is that you are you're getting professional grade data extraction without having to

04:01.330 --> 04:02.810
build these tools yourself.

04:03.530 --> 04:05.210
And some of the scrapers are free.

04:07.090 --> 04:08.970
What's impressive is the scale here.

04:09.810 --> 04:17.210
There are over 5000 actors, so different scrapers available across categories like social media agents,

04:17.250 --> 04:18.170
lead generation.

04:18.850 --> 04:23.290
So it's really comprehensive for almost any data extraction task you can think of.

04:24.850 --> 04:32.170
And the beauty of this new Apify MCP server is that it lets you use all these different scrapers through

04:32.170 --> 04:36.250
one credential, instead of managing separate keys for each service.

04:37.090 --> 04:41.890
So just a few months ago, if you wanted to call different services, you would typically need to send

04:41.890 --> 04:44.090
requests with API keys.

04:44.170 --> 04:49.410
So each key was your credential that opened access to that specific service.

04:49.890 --> 04:55.210
So you'd have one key for Instagram scraping, another for LinkedIn, another for Google Maps, and

04:55.210 --> 04:55.730
so on.

04:57.530 --> 05:04.330
So each key gives you access to one specific type of data or service by using MCP servers.

05:04.370 --> 05:10.130
One key or one set of credentials can access multiple services at the same time.

05:11.210 --> 05:14.930
So before you'd have to create separate credentials to scrape Instagram.

05:15.210 --> 05:16.690
Then another set for Amazon.

05:16.930 --> 05:19.210
Another for X, another for Facebook.

05:19.330 --> 05:22.170
Now you can use the same credential for all of them.

05:23.970 --> 05:29.490
So on this diagram you can see that on the left you have MCP clients.

05:29.490 --> 05:35.130
So these are applications like cloud desktop cursor or other AI tools.

05:35.610 --> 05:41.090
In the middle you have the API MCP server which acts as a central hub.

05:41.410 --> 05:46.170
And on the right you can see all the different platforms it can connect to.

05:46.370 --> 05:50.810
So Instagram, TikTok X and thousands of other services and platforms.

05:51.210 --> 05:56.770
The key difference is that instead of each client needing separate connection to each platform.

05:57.090 --> 06:00.410
Everything goes through this central MCP server.

06:00.650 --> 06:06.970
So when you make a request in Cloud Desktop, it can automatically access Instagram for posts, TikTok

06:06.970 --> 06:11.770
for videos, X for tweets, and all through that single connection.

06:12.170 --> 06:17.370
So with MCP servers, you can make one request that has different conditions.

06:18.090 --> 06:22.770
So with MCP servers, you can make one request that has different conditions.

06:23.210 --> 06:28.170
And it's intelligent enough to figure out which services to use and in what order.

06:28.330 --> 06:35.330
It automatically handles all the different platforms and executes the actions that makes the most sense.

06:36.050 --> 06:40.450
So for actually using MCP servers, I'm working with Cloud Desktop here.

06:40.930 --> 06:47.930
So thanks to that I can have conversations with large language model like Cloud Sonnet or Cloud Opus

06:48.290 --> 06:50.930
that has access to all these different scrapers.

06:51.130 --> 06:52.290
You have been looking at.

06:53.130 --> 07:00.690
So when I click on this drop down, you can see that here I have access to a file server and I have

07:00.690 --> 07:02.810
access to all these different options.

07:04.450 --> 07:10.090
And what's really cool is that some of these tools will actually help me search Apify to discover what

07:10.130 --> 07:12.930
services exist for whatever request I have.

07:13.490 --> 07:19.170
So if you haven't set up a specific scraper yet, this system can help me locate the best ones on the

07:19.170 --> 07:19.770
platform.

07:20.370 --> 07:23.570
It can find scrapers based on reviews and pricing.

07:24.810 --> 07:27.010
So setting this up is actually very simple.

07:27.290 --> 07:32.610
We need to edit a configuration file in Cloud desktop app as we did in our previous lessons.

07:34.130 --> 07:39.450
So go to Cloud Desktop app and in the top menu click on settings.

07:43.810 --> 07:45.050
Then Developer settings.

07:46.330 --> 07:50.890
So when I click on Edit config it opens up a folder.

07:50.890 --> 07:52.690
And then when I click on this file.

07:53.130 --> 07:55.650
So cloud desktop config JSON.

07:58.930 --> 08:01.850
It shows me all the credentials I currently have set up.

08:02.530 --> 08:07.410
Now if you haven't used Cloud desktop app before, this will be a blank file which is actually easy

08:07.410 --> 08:08.130
to work with.

08:08.250 --> 08:12.690
In my case, you can see I have a different services already configured.

08:14.170 --> 08:23.810
So when I scroll down you can see that I have an SMTP server with some configuration details and the

08:23.850 --> 08:25.050
actual service name.

08:28.330 --> 08:29.890
Apify SMTP server.

08:30.490 --> 08:36.730
So below that is the list of different scrapers that I want to have available out of the box.

08:38.290 --> 08:47.290
So I can use YouTube comment scraper TikTok scraper, Instagram scraper, YouTube transcript scraper,

08:48.290 --> 08:51.650
tweet scraper, Facebook ad scraper, and so on.

08:53.730 --> 08:55.810
So adding another scraper is really simple.

08:55.810 --> 08:57.770
Let's say I want to add something new.

08:58.330 --> 09:00.290
I can go back to Apify store.

09:01.610 --> 09:06.650
First of all, if you are new to Apify, please use the link from the resources section of this lesson.

09:07.370 --> 09:11.410
Create your account and sign in by using this link.

09:11.450 --> 09:13.970
You will get $5 in free credits to use.

09:14.770 --> 09:16.210
Then go to Shopify store.

09:19.530 --> 09:26.490
And here you can browse by category so you can find scrapers for social media, for e-commerce, SEO

09:26.490 --> 09:27.450
tools and so on.

09:28.090 --> 09:32.090
Or you can search by specific platform, for example, LinkedIn.

09:33.490 --> 09:37.090
Let's say I want to scrape LinkedIn profiles with emails.

09:38.330 --> 09:41.130
You can also set up filters for example popular.

09:42.090 --> 09:43.370
So I'm going to select this one.

09:46.690 --> 09:48.970
Obviously we want to double check that.

09:48.970 --> 09:51.330
It's going to give you the data you are expecting.

09:52.050 --> 09:56.930
So in order to do this, you can run this and see what kind of results you get.

09:57.770 --> 10:02.490
So assuming you find it useful, you would need to look at the price too.

10:02.850 --> 10:03.770
So in that case.

10:04.250 --> 10:09.690
So for this particular scraper you would need to pay $10 for 1000 results.

10:19.530 --> 10:24.370
I'm not going to open this LinkedIn profile for privacy reasons.

10:25.690 --> 10:27.810
Let's say I want to scrape Facebook ads.

10:28.610 --> 10:33.090
So I'm going to search for actor type Facebook.

10:37.170 --> 10:37.890
And here we go.

10:38.570 --> 10:45.050
You can add this scraper to analyze your competitors advertising strategies or identify trending products

10:45.090 --> 10:45.730
in niche.

10:46.210 --> 10:48.690
So you can find a lot of scrapers too.

10:49.130 --> 10:51.420
And now in order to add the scraper.

10:51.420 --> 11:01.220
So this tool to our MCP server, all we need to do is literally copying this name and adding to our

11:01.660 --> 11:02.180
file.

11:03.380 --> 11:09.260
So we are going to click on this copy button and paste it back into our configuration file.

11:10.500 --> 11:12.020
So the process is very simple.

11:12.420 --> 11:23.580
In our actor MCP server connection and the actors, just after any of the actors add a new name of the

11:23.580 --> 11:24.220
scraper.

11:29.820 --> 11:31.460
When I check my cloud desktop app.

11:40.220 --> 11:41.540
And click on this drop down.

11:41.540 --> 11:43.020
So my tools library.

11:46.060 --> 11:49.220
You can see that in the Apify MCP SMTP server.

11:49.500 --> 11:52.260
I have access to 28 tools.

11:53.300 --> 12:00.380
And now when I restart the cloud and add the name of the new scraper and comma.

12:02.420 --> 12:04.180
And then save the file.

12:09.220 --> 12:12.940
Now when I click on this drop down to see my tools library.

12:15.540 --> 12:18.780
You can see that in the SMTP server.

12:19.500 --> 12:23.660
Now I've got access to 29 tools not 28.

12:24.420 --> 12:30.060
And when I type Facebook I have access to this specific scraper.

12:30.060 --> 12:32.500
So apify Facebook ad scraper.

12:34.700 --> 12:36.340
So this is scraper we just added.

12:37.580 --> 12:45.180
If for whatever reason you want to remove scraper you can either toggle it off as simple as that, or

12:45.180 --> 12:48.700
go back to that configuration file and delete it from the list.

12:49.220 --> 12:51.380
Just remember to remove the comma as well.

12:52.020 --> 12:54.740
Now let's actually see our scrapers in action.

12:55.020 --> 12:58.460
So first let me test the Google Maps scraper.

12:58.460 --> 13:11.060
So I will ask it to find ten digital marketing agencies in San Francisco.

13:15.060 --> 13:15.940
With their.

13:19.980 --> 13:21.140
Contact details.

13:23.380 --> 13:33.140
Ratings and service areas for potential partnership.

13:36.540 --> 13:37.540
And hit send.

13:38.380 --> 13:43.580
Actually I'm going to add use Google maps.

13:45.940 --> 13:53.620
To find I'm being specific about using Google Maps because this could be this could technically be done

13:53.660 --> 13:59.540
with web search, but Google Maps give us more accurate and more richer business data.

14:00.140 --> 14:08.460
So by default, cloud language model could use web search instead of using this tool in our SMTP server.

14:08.860 --> 14:10.300
So I'm going to hit send.

14:15.500 --> 14:23.540
So if you find it's not using the tool you want you can explicit and say use for example Apify Google

14:23.540 --> 14:25.300
Maps scraper tool.

14:25.740 --> 14:27.780
And it will prioritize that specific tool.

14:28.180 --> 14:33.500
Or you can just toggle off the default web search tool in Cloud desktop app.

14:35.340 --> 14:38.900
So now cloud asks us about the permission to use this tool.

14:39.260 --> 14:41.900
So I'm going to allow always.

14:43.980 --> 14:45.180
Here you can see the request.

14:49.060 --> 14:50.100
Let's wait a moment.

14:51.380 --> 14:54.700
Now it's using Compass Cloud Google places tool.

14:58.220 --> 15:02.260
So the tool we have connected to our MCP server.

15:04.740 --> 15:05.860
Again you can find it here.

15:11.100 --> 15:14.180
Now they're asking us about using another tool.

15:15.020 --> 15:17.140
I'm going to allow always.

15:21.700 --> 15:24.820
I'll help you find digital marketing agency in San Francisco using Google Maps data.

15:24.860 --> 15:29.660
Let me search for agencies and extract the contact details, ratings and service information.

15:30.140 --> 15:34.820
Now let me get more details from the complete data set to give you comprehensive information about all

15:35.220 --> 15:36.020
agencies.

15:36.420 --> 15:36.900
Perfect.

15:36.900 --> 15:41.100
I found ten digital marketing agencies in San Francisco with comprehensive details.

15:41.500 --> 15:46.340
Here is a detailed overview of each agency for potential partnership consideration.

15:48.060 --> 15:50.380
Top digital marketing agencies in San Francisco.

15:50.700 --> 15:57.140
We have a rating of address, website, email or paying our services specialties.

16:06.300 --> 16:06.860
Awesome.

16:16.260 --> 16:18.100
For partnership consideration.

16:18.140 --> 16:19.660
Top performers by reviews.

16:21.300 --> 16:22.140
Service areas.

16:22.140 --> 16:23.300
Common specialties.

16:23.860 --> 16:25.340
Contact recommendations.

16:30.020 --> 16:32.740
So now we can actually just check one of the agency.

16:50.460 --> 16:51.180
Regret.

16:55.220 --> 16:56.420
For the next test.

16:56.460 --> 16:58.020
Let me try TikTok scraper.

16:58.260 --> 16:59.300
Now you search for.

17:04.420 --> 17:05.060
TikTok.

17:08.780 --> 17:09.500
Videos.

17:11.700 --> 17:12.540
With AI.

17:14.380 --> 17:16.220
AI ML.

17:16.220 --> 17:16.820
Scraper.

17:17.100 --> 17:17.860
The hashtag.

17:22.380 --> 17:23.460
Could you find me?

17:25.780 --> 17:26.580
Pardon me.

17:27.460 --> 17:32.740
Five viral TikTok videos with that hashtag.

17:34.700 --> 17:34.940
Okay.

17:34.940 --> 17:35.460
Let's see.

17:35.580 --> 17:44.310
So this is a trending hashtag where people use AI to create viral ASMR content with with Google's VR

17:44.310 --> 17:46.710
three sedans or link two.

17:46.750 --> 17:48.390
So text to video models.

17:49.870 --> 17:50.310
Right.

17:54.630 --> 17:58.350
Now it's using clockworks TikTok scraper.

18:00.670 --> 18:07.230
So I limit it to five videos and the system goes through multiple accounts, view counts like comments

18:07.230 --> 18:07.870
and so on.

18:09.550 --> 18:13.470
Let me get more results to find better viral videos with higher engagement.

18:16.830 --> 18:19.310
I'll help you find viral TikTok videos.

18:20.750 --> 18:24.270
Let me search for TikTok videos using relevant hashtags and search terms.

18:24.870 --> 18:29.150
Let me get more results to find better viral videos with higher engagement.

18:29.550 --> 18:30.030
Perfect!

18:30.030 --> 18:37.950
I found some great viral TikTok videos with with eye ASMR content based on the search results.

18:38.150 --> 18:43.150
Here are the five most viral TikTok videos with the with this hashtag.

18:44.630 --> 18:45.070
Okay.

18:49.630 --> 18:53.710
So as you can see, each results includes direct links to watch the videos.

18:57.150 --> 19:03.710
We've got likes, shares, views, comments, description, duration and why it's vital.

19:04.430 --> 19:05.950
All right let's check this one.

19:17.150 --> 19:17.550
Okay.

19:19.870 --> 19:20.550
And this one.

19:40.990 --> 19:41.630
All right.

19:42.030 --> 19:50.270
So this kind of social media system would normally require specialized tools or manual research.

19:50.990 --> 19:53.390
Now for Instagram I will ask you to use.

19:55.750 --> 19:56.950
Instagram scraper.

20:02.630 --> 20:03.430
To analyze.

20:05.550 --> 20:07.950
The last 20 posts.

20:11.510 --> 20:12.670
From this account.

20:15.910 --> 20:20.870
I'm going to provide it with a the name of this account.

20:21.310 --> 20:22.270
Interesting AF.

20:23.070 --> 20:24.430
So we want to scrape this account.

20:29.270 --> 20:32.230
And also we want it to identify.

20:35.990 --> 20:36.950
The most.

20:40.910 --> 20:42.030
Engaging posts.

20:46.190 --> 20:47.390
The most engaging.

20:50.110 --> 20:50.710
Topics.

20:53.910 --> 20:56.030
And posting patterns.

20:59.190 --> 21:00.150
And hit send.

21:04.030 --> 21:10.030
So it automatically selects the Instagram scraper and receives the data.

21:13.070 --> 21:18.630
I'll help you analyze the Instagram account using Instagram Scraper to examine the last 20 posts and

21:18.630 --> 21:22.270
identify engaging posts and topics and posting patterns.

21:25.270 --> 21:28.030
So it's using Shopify Instagram Scraper.

21:32.870 --> 21:37.350
In the request, we pass the URL and we also limited the results to 20.

21:39.670 --> 21:40.790
Right thinking now.

21:48.630 --> 21:49.790
Scrape Instagram data.

21:49.790 --> 21:51.870
Analyze engagement and posting patterns.

21:57.270 --> 22:02.630
Now let me get the complete data set to analyze all the posts and their engagement patterns.

22:03.870 --> 22:06.910
Now it's using get dataset items to.

22:13.150 --> 22:14.510
Analyze Instagram posts.

22:14.670 --> 22:15.150
Perfect!

22:15.150 --> 22:18.910
I successfully scraped 19 posts from this account.

22:19.830 --> 22:20.150
Perfect.

22:20.150 --> 22:21.030
Let's see the results.

22:24.670 --> 22:26.790
A Gallery engagement report.

22:28.430 --> 22:34.470
The analysis of the last 19 posts from this strategy focused on AI generated content.

22:35.750 --> 22:37.550
Top performing posts by engagement.

22:41.790 --> 22:42.910
We got the dates.

22:43.910 --> 22:44.790
Captions.

22:45.430 --> 22:46.030
Comments.

22:46.030 --> 22:46.590
Likes.

22:49.950 --> 22:52.030
Most engaging content categories.

22:55.870 --> 22:57.430
Posting patterns and strategy.

22:59.430 --> 23:00.790
Content source strategy.

23:02.390 --> 23:03.550
Key success factors.

23:06.630 --> 23:08.550
Even recommendations for optimization.

23:09.430 --> 23:11.070
So engagement rate analysis.

23:11.110 --> 23:14.110
Average likes over 2000 likes.

23:14.430 --> 23:15.510
Average comments.

23:15.750 --> 23:17.790
Video performance interactive content.

23:18.630 --> 23:19.190
Look at that.

23:19.190 --> 23:20.310
It's very comprehensive.

23:20.990 --> 23:22.350
Most engaging topics.

23:24.510 --> 23:27.150
Content strategy breakdown top performing patterns.

23:27.910 --> 23:29.390
So we received what we asked for.

23:29.910 --> 23:30.710
That's amazing.

23:31.070 --> 23:39.110
So this kind of competitive social media analysis would normally require A lot of work and data collection.

23:39.670 --> 23:41.950
Now let me test the YouTube transcript scraper.

23:42.270 --> 23:48.030
So I'll ask it to extract the full transcript.

23:52.670 --> 23:53.670
From this video.

23:55.710 --> 23:58.510
And I'm going to provide the link to.

24:00.790 --> 24:04.070
The last video published by Y-combinator.

24:04.790 --> 24:08.070
So you want it to scrape the transcript from this video.

24:09.030 --> 24:12.070
So the Andrej Karpathy Talk software is changing again.

24:19.670 --> 24:20.550
And create.

24:23.510 --> 24:24.550
A structured.

24:27.990 --> 24:28.870
Summary.

24:31.110 --> 24:36.200
With the main steps and key.

24:38.480 --> 24:42.240
Takeaways organized by topic.

24:48.120 --> 24:49.040
They presented.

24:52.480 --> 24:53.440
As a visual.

24:56.480 --> 24:57.160
Mind map.

24:57.920 --> 25:00.400
By the way, I strongly recommend you watching this video.

25:00.840 --> 25:03.960
Andrej Karpathy is a genius, actually.

25:04.000 --> 25:08.360
For this example, let's use the extended thinking feature.

25:11.640 --> 25:14.040
So it will improve the scraping process.

25:14.760 --> 25:17.080
And now you can see the decision making process.

25:18.160 --> 25:24.160
So it determines that it needs the YouTube transcript scraper from Apify and automatically structures

25:24.160 --> 25:25.240
the API request.

25:25.720 --> 25:31.840
Something you would normally have to configure manually in tools like N810 or Make.com.

25:34.880 --> 25:39.640
So I will help you extract the transcript from that YouTube video and create a structured summary with

25:39.680 --> 25:40.680
a visual mind map.

25:40.880 --> 25:43.000
Let me start by extracting the transcript.

25:45.080 --> 25:48.200
As you can see, it's using YouTube transcript scraper.

25:48.720 --> 25:52.240
Now let me get the full transcript since it was truncated.

25:52.840 --> 25:53.360
Right.

25:55.800 --> 25:56.160
Perfect.

25:56.160 --> 25:58.920
I have extracted the full transcript for Andrej Karpathy's talk.

26:00.400 --> 26:05.080
We have the main thesis topics and takeaways.

26:06.920 --> 26:09.040
We've got our alarms and operating systems.

26:10.320 --> 26:12.120
Now it's creating the mind map for us.

26:14.200 --> 26:16.200
Let's wait for our mind map.

26:27.360 --> 26:28.040
All right.

26:28.080 --> 26:28.920
That's impressive.

26:31.640 --> 26:32.480
That's amazing.

26:36.120 --> 26:41.680
So as you can see the complete transcript organized summaries.

26:42.360 --> 26:43.760
And now when I click publish.

26:49.280 --> 26:49.680
Link.

26:52.920 --> 26:54.200
And pass it to my browser.

26:58.520 --> 27:02.320
I have this beautiful mind map from Andrej Karpathy talk.

27:02.680 --> 27:04.040
Software is changing again.

27:05.760 --> 27:08.600
So we have organized summaries and structured breakdowns.

27:09.120 --> 27:14.760
So I asked it to create a visual mind map showing how all the concepts connect to each other.

27:16.080 --> 27:25.000
And from one URL and one request, we get this comprehensive analysis presented in the mind map that

27:25.000 --> 27:28.480
would typically require multiple separate tools and automations.

27:29.160 --> 27:30.640
That's just mind blowing.

27:31.600 --> 27:32.520
I really like it.

27:35.080 --> 27:43.680
So we've got six main topics everyone as a programmers vibe coding with natural language, building

27:43.680 --> 27:46.760
for agents, right.

27:47.400 --> 27:51.520
Software paradigms, llms as operating systems.

27:51.960 --> 27:53.280
That's an interesting topic.

27:54.080 --> 27:57.600
LLM psychology and partial autonomy apps.

27:58.760 --> 28:02.160
I really enjoyed this talk and I highly recommend you watching it.

28:03.400 --> 28:07.800
So here we also have a summary of the process and the topics covered.

28:08.040 --> 28:08.960
Main takeaways.

28:09.880 --> 28:15.040
So in the text format now let's test something for recruitment.

28:15.320 --> 28:16.600
So I'll ask it to.

28:20.120 --> 28:22.520
Use LinkedIn.

28:25.800 --> 28:26.520
Scraper.

28:28.560 --> 28:29.280
Find.

28:33.880 --> 28:37.600
A automation specialist.

28:42.280 --> 28:43.600
In the Boston area.

28:47.440 --> 28:49.400
Who have experience?

28:53.960 --> 28:58.360
With tools like N810.

29:00.160 --> 29:04.640
Maybe to find five automation specialists.

29:06.960 --> 29:08.000
And let's see.

29:11.080 --> 29:12.400
If I can click the tools.

29:14.720 --> 29:17.840
It's asking for permission to use the scraper.

29:18.720 --> 29:20.080
I'm going to allow it once.

29:25.040 --> 29:29.360
I help you find a automation specialist in the Boston area with experience.

29:30.200 --> 29:30.520
All right.

29:30.560 --> 29:33.800
Thinking Now it's asking for permission again.

29:33.800 --> 29:37.440
So I'm going to allow always this time.

29:41.400 --> 29:46.240
Let me try a broader search with more general automation and AI related terms to find relevant professionals.

29:49.440 --> 29:49.960
Okay.

29:54.840 --> 29:59.680
Let me try simple search approach with just basic automation terms and see if we get some results.

30:05.520 --> 30:05.880
Great.

30:05.880 --> 30:07.080
I found some relevant profiles.

30:07.080 --> 30:13.680
Let me get more details from the search results and do another search to find more automation specialists.

30:14.600 --> 30:18.080
Now let me run another search to find more automation specialists with different search terms.

30:18.840 --> 30:20.160
Alright, we have to be patient here.

30:20.880 --> 30:21.200
Perfect.

30:21.240 --> 30:23.680
Now I have some excellent results from the second search.

30:25.280 --> 30:28.040
Let me get full details from this latest search.

30:28.720 --> 30:29.000
All right.

30:29.000 --> 30:31.880
So we have some top AI automation specialists in Boston area.

30:34.240 --> 30:34.560
All right.

30:34.560 --> 30:35.040
Great.

30:35.880 --> 30:41.520
We have title company location, skills experience link to LinkedIn profile.

30:41.960 --> 30:44.040
And why is it perfect for ten.

30:48.360 --> 30:50.080
Best matches for experience.

30:51.080 --> 30:51.400
All right.

30:51.400 --> 30:51.800
Awesome.

30:52.360 --> 30:56.080
Now, um, finally I will combine data scraping with analysis.

30:56.520 --> 30:57.920
So I'm going to ask.

31:00.360 --> 31:01.280
Use Twitter.

31:08.200 --> 31:09.200
With the scraper.

31:16.320 --> 31:20.720
To scrape the last 50.

31:24.640 --> 31:28.960
Or maybe 20 viral posts.

31:30.730 --> 31:31.370
from.

31:33.410 --> 31:42.410
Tech tech industry industry leader and analyze.

31:45.250 --> 31:46.210
The teams.

31:48.010 --> 31:51.330
Predict the next trends.

31:53.450 --> 32:00.490
Let me start scraping posts from OpenAI CEO, since it has been very influential in AI trends.

32:00.490 --> 32:03.170
If you would prefer a different leader, Elon Musk.

32:03.490 --> 32:04.450
Jensen Huang.

32:04.850 --> 32:05.810
Satya Nadella.

32:06.610 --> 32:07.290
Let me know.

32:10.410 --> 32:12.890
You can see what he's doing now step by step.

32:13.330 --> 32:13.570
Yes.

32:13.570 --> 32:20.170
So the results include predictions based on the based on patterns in the content.

32:20.370 --> 32:27.090
So based on the viral tweets from tech leaders such as Sam Altman.

32:30.370 --> 32:31.290
Satya Nadella.

32:31.450 --> 32:32.250
Elon Musk.

32:33.890 --> 32:35.050
Her team analysis.

32:37.690 --> 32:40.610
Trend predictions for 2025 2026.

32:41.290 --> 32:42.370
Autonomous agents.

32:42.370 --> 32:45.810
Mainstream consumer brain computer interface.

32:45.810 --> 32:46.490
Quantum AI.

32:46.530 --> 32:47.650
Hybrid systems.

32:48.050 --> 32:48.410
Retail.

32:48.410 --> 32:49.250
Multimodal AI.

32:49.730 --> 32:51.210
Native infrastructure.

32:51.490 --> 32:53.010
Personalized AI companions.

32:53.530 --> 32:54.730
Predicted timeline.

32:57.610 --> 32:59.090
Actionable insights.

32:59.650 --> 33:00.730
Investment opportunities.

33:00.730 --> 33:01.570
Business strategy.

33:01.570 --> 33:02.730
Technology focus.

33:03.490 --> 33:04.730
Implementation guide.

33:04.730 --> 33:06.170
How to run this analysis.

33:06.370 --> 33:13.370
So you got this nice tech trend analysis dashboard to analyze viral posts from tech industry leaders

33:13.370 --> 33:15.290
to predict future trends.

33:16.090 --> 33:19.850
But as you can see here, it didn't use that Twitter scraper.

33:19.890 --> 33:25.170
Since I haven't bought the credits, it fell back to the default search feature.

33:25.650 --> 33:31.170
And these are just examples of how powerful this apify SMTP server can be.

33:31.570 --> 33:37.410
Of course, to set this up yourself, I will include the complete configuration schema in the resources

33:37.410 --> 33:44.130
section of this lesson, so you can import it directly into your cloud desktop app configuration file,

33:44.130 --> 33:49.450
and have some of the scrapers offered by Apify SMTP server ready to use.

33:49.730 --> 33:57.330
These tools show the potential because instead of juggling multiple scraping tools and API keys, you

33:57.370 --> 34:02.290
get intelligent data extraction that adapts to your specific needs.

34:02.290 --> 34:05.690
So you can combine multiple scrapers in a single request.

34:05.930 --> 34:13.330
And the LLM will act like an AI agent that intelligently decides which tools to use and in what order.

34:13.930 --> 34:16.890
Plus, you are not limited to just data extraction.

34:17.210 --> 34:23.250
You can add analysis, visualization, and even generate insights in one request.

34:23.490 --> 34:24.730
I hope this was helpful.

34:24.970 --> 34:26.970
Thanks for watching and see you in the next one.