N8N Full Course: Build AI Automations in 2026 (For Beginners)

0:00Welcome to the most comprehensive N8N

0:01course, where I take you from a real

0:03beginner in AI automation [music] to

0:05someone who's actually able to build

0:07automations and AI agents for themselves

0:10or for other businesses. And in case we

0:11haven't met, my name is Michele, and

0:13over the past 12 months, we've helped

0:15over 40 businesses implement AI and

0:17automations and taught over 20,000

0:19people in the process. All starting with

0:21zero technical knowledge. So, we're

0:22going to go through everything from

0:23understanding

0:24>> [music]

0:24>> what is AI automation to setting up your

0:26own N8N account, understanding how to

0:29build stronger and faster, more AI

0:31automations and AI agents, and finally,

0:34actually building systems step-by-step.

0:36[music]

0:36By the end, you'll be an AI automations

0:38and AI agents expert, and you'll know

0:40exactly how to build systems that drive

0:42real business value. So, the course

0:44itself is split into six different

0:46modules. Module one is getting started

0:48with N8N foundations. We'll go through

0:50exactly what automation even is to

0:52understanding how you can set up

0:53everything within N8N and build your

0:55first automation. Module two is [music]

0:57core N8N foundations, using automation

0:59logic and APIs, and walking you through

1:01exactly how automation really works in

1:03the backend. Module three is AI agent

1:05fundamentals, and really understanding

1:07how AI agents work. Module four is

1:09building smarter AI systems, using

1:11frameworks and prompts, which allow our

1:13automations and AI agents to make sure

1:15that everything that we build is

1:16actually scalable, tangible, and can

1:18actually be implemented into a business.

1:20Module five is real-world AI agent

1:22projects, end-to-end builds, showing you

1:24exactly different use cases within a

1:26business that you can implement [music]

1:28AI agents, and showing you how I build

1:30it step-by-step. And module six is

1:32real-world AI workflow projects,

1:34end-to-end builds, [music] where I show

1:35you exactly how we now not only just use

1:38AI agents, but use entire workflows to

1:40be able to implement into a business,

1:42and showing you how I build different

1:43use cases step-by-step. By the way, if

1:45you want access to all the resources

1:47that we're going to be talking about in

1:49this video, plus being part of a

1:51community of over 3,000 people who are

1:53actually building automations every

1:54single day and growing their own

1:56businesses, check the second link down

1:57below. I'm going to leave all the

1:59timestamps down below in case you want

2:01to skip through some modules or go to

2:03the direct part, or even come back to

2:04this video uh because I know it's really

2:06long. Um so, without further ado, let's

2:09dive in.

2:12All right, before we jump into N8N,

2:14let's talk about what is automation.

2:16Automation is the process of using

2:17technology to do tasks without you.

2:20Whether it's AI agents, whether it's

2:21workflows, whether it's all this stuff,

2:23that's exactly what it is. It's using

2:25softwares

2:26to be able to do things without you.

2:28Now, every day, we all perform many

2:30regular ad-hoc tasks. Ad-hoc just means

2:32that it's repetitive. Whether you are a

2:33working professional, whether you are a

2:35business owner, whether that's in sales,

2:37marketing, operations, finance, or

2:39business management, we all have

2:41different parts of the business that we

2:43have to do because that's what drives

2:44the business forward. But the problem

2:46is, these tasks are repetitive, so they

2:49happen every single time. They're

2:50no-to-low brainer, so they're not that

2:52hard to do.

2:53They're low value addition, so you do

2:55them, but they don't add that much value

2:57into your day. And most importantly,

2:58they're effort and time consuming. So,

3:00you do them, they take a ton of time,

3:02you don't want to do them, they are

3:03repetitive, they're very easy, and so

3:05the solution is automation. Now, you

3:07might not be fully conscious of it, but

3:09automation actually plays a big role in

3:11our everyday lives. For example, when we

3:13have a meeting in our school community,

3:14so we have two weekly calls. One is a

3:16weekly huddle, the other one is a vibe

3:17coding session. When the meeting is 1

3:19hour away, then it sends a notification

3:21over email to my school members. Or

3:23Netflix. When you watch a certain type

3:26of movie, then Netflix recognizes that,

3:28so it says, "Hey, he watched this movie,

3:29that means he must like something else,

3:31right?" And it gives you a

3:32recommendation. That is automation,

3:34because it automates the process of you

3:36having to find the movie that you like

3:38in yourself.

3:39Or a coffee machine. And you might be

3:40asking yourself, how does a coffee

3:42machine even play a role in this? Well,

3:43just hear me out. Whenever a coffee

3:45machine finishes making the coffee, then

3:47the timer finishes. And lastly, ChatGPT,

3:50right? Because when you give a prompt to

3:52ChatGPT, then it gives you the output,

3:54right? Now, ChatGPT is a bit different

3:56just because it's non-deterministic. So,

3:59non-deterministic means that one series

4:01of input. So, you can say, "Hey, can you

4:03do this for me?" And it does something

4:04else. You can say the exact same thing,

4:06and it might do it differently next

4:07time, right? Because you can't determine

4:09the output that you get. But at the end

4:10of the day, everything is automation.

4:12Now, I don't know if you noticed the way

4:13that I was actually explaining these

4:15concepts, but we actually use a very

4:17simple principle, which is if this, then

4:20that. So, if we have X, then it goes to

4:22either A and B. If it's B, it goes to

4:24either E and F. If it's A, then it goes

4:26to either C and D. And that is exactly

4:28how automation works. Everything and

4:30everything is built on logic, right?

4:32Whether you're building code, whether

4:33you're using automations, whether you're

4:34using N8N, make.com, Zapier, Power

4:37Automate, all these tools, they're all

4:38built on logic. If this happens or when

4:41this happens, then you do this. And the

4:43question then becomes, what can

4:44automation do for you? Well, as a

4:46business owner or an individual, it

4:49increases productivity and efficiency,

4:50which means that you can do more with

4:52less, more leverage. You save time to do

4:54other things, right? Because automation

4:56for a business means that they can now

4:58focus on other areas of the business.

5:00Has an increase of availability, cuz

5:02automation doesn't need holidays, they

5:04don't need any emotional support,

5:05they're always there working 24/7. You

5:07can scale with less resources. And by

5:09the way, we've come to a point where

5:10automation allows us to be able to build

5:12a one-person company, as you might have

5:14heard. Uh but it's all because of the

5:16softwares and technologies that is out

5:17there. All because of automation.

5:20And it's more robust and less errors,

5:22especially when it comes to the

5:23repetitive tasks that you have to do

5:24every single time. Cuz if you are

5:26anything like me, when I do something

5:28over and over again, I get tired, and my

5:31quality of output goes down, right? And

5:33so, with automation, it's consistent.

5:35And with that said, let's jump into N8N.

5:38All right, so this right here is N8N.

5:40Now, all you have to do to get to this

5:42page is go to n8n.io, and you'll have

5:45this dashboard. Now, as you can see

5:46here, it's a flexible AI workflow

5:48automation for technical teams that

5:50allows us to be able to build

5:52automations without needing to know how

5:54to code. Now, the automation can vary,

5:56right? Depending on what you're using

5:57it. IT, security, dev,

6:00sales, you can do a bunch of stuff

6:01within this exact platform, and it

6:03allows us to have so much flexibility

6:05around what we build. As well as

6:06plugging it to over 500 integrations,

6:08which is a bit lower than other

6:10platforms, but it still allows us to be

6:12able to speak to different softwares all

6:14together. And also, it has the ability

6:15for us to code when you need it, the UI

6:18when you don't, which means that you can

6:20go either way, which is why I mentioned

6:21that it's flexible in what we do,

6:23because it's great for two kinds of

6:24people. All right, so if I go to pricing

6:26right here, I can see how much we

6:27actually have to spend to be using N8N.

6:30Now, N8N does not have a free plan on

6:32the cloud version, right? Cloud version.

6:35Now, if you're a beginner to N8N, you

6:37have probably heard about self-hosting

6:39versus cloud hosting. Cloud hosting just

6:41means that you just use N8N within the

6:43cloud. Literally just here. On Chrome,

6:46on Safari, whatever it is. But we do

6:48have the option to self-host, which

6:50means that we create our own server, and

6:52we self-hosted in our own server, so we

6:54don't have to pay. Now, don't worry, I'm

6:56going to show you exactly how to

6:57self-host as well in the next few

6:59videos, so you'll have that, too. Now,

7:01for the cloud hosting version, which is

7:02the one that most people use, we don't

7:04have a free plan, but we have a starter

7:06plan, pro, business, and enterprise.

7:09Now, most people use the 24 euro plan,

7:11which is, I believe, about 30 bucks. And

7:14the way that N8N actually prices us is

7:16through workflow executions, which I'll

7:18speak about in just a second. But we do

7:19have the ability for us to start a free

7:22trial with no credit card required,

7:24which allows us to test N8N for about 14

7:26days before we have to upgrade to a paid

7:28plan.

7:29And here, we have one shared project,

7:32five concurrent executions, unlimited

7:34users, 50 AI workflow build credits, and

7:37forum support, which is more than enough

7:39for what we need. Now, if you are

7:40building N8N for a bigger company and

7:42you want more workflow executions, then

7:44you're most likely going to be using the

7:46pro plan, which is 60 euros a month.

7:48Now, honestly, this is not super cheap,

7:50but when you think about the concept of

7:51how they actually price their

7:52automations, it is much better than any

7:55other platform that I've seen before.

7:56And same thing here, but it has a bit

7:57more stuff when it comes to the actual

7:59plan. All right, so for you, just press

8:01start free trial right here, then we can

8:03start filling out the form.

8:05And then you can press start a free

8:0714-day trial.

8:09Go through the onboarding questionnaire

8:11right here,

8:12where they ask you a few questions just

8:14so they get to know you a little bit

8:15more, okay?

8:18Podcast, submit. And you can also invite

8:20members into your workspace. Let's just

8:22press skip. And now our workspace is

8:23starting up. Now, I do really like the

8:25onboarding um sequence that N8N has.

8:28It's really, really quick, and it's

8:29really, really easy to follow. Now,

8:30because you're watching this video,

8:31there's no need to press this video

8:33right here, cuz I'll show you everything

8:35that you have to know. Uh but it is a

8:36pretty good starting video as well. And

8:38this guy is a legend.

8:40>> [laughter]

8:40>> All right, once the workspace is ready,

8:42just press start automating, which will

8:43take you to

8:45a dashboard that looks like this. So,

8:47welcome to N8N. Now, on the top, you get

8:49to see that you have 14 days left on

8:51your free trial, and you have 8,000

8:53executions, which is again the way that

8:54N8N charges you. And then on the top,

8:56you have upgrade now, which is a button

8:58to upgrade. I'm going to press X, and

8:59now we have this page. All right, so on

9:01the top right here, we have a dashboard

9:03that allows us to visualize the numbers.

9:05So, I can actually press inside each one

9:07and be able to see

9:08the different graphs. Now, production

9:11executions just mean that, "Hey, how

9:12many executions?" So, how many times did

9:14our automation run every single 7 days?

9:18How many times did the automation run,

9:20but it didn't work? Then out of all the

9:22executions that didn't run in the past 7

9:23days, how many of them did not work,

9:25which is this percentage right here? And

9:27then, how much time did you save, or how

9:28much time did the workspace save with

9:30these automations in place, which is

9:32good to have, but at the same time, it

9:33isn't super accurate, cuz to understand

9:35how much time you saved for a company or

9:37for yourself, N8N needs to know how much

9:39time did you take in the first place to

9:41do it, right? Because then you're saving

9:43them that much time. And lastly, run

9:45time. So, for how long are the

9:46automation running? Now, right here, and

9:48on the top as well, you have the ability

9:49to create a workflow, and this is how

9:51you start from scratch. This is how you

9:52start building. Now, before we get to

9:53that, we have workflows here, then we

9:55have credentials right here, which is

9:57the ability for us to be able to add

9:59credentials for the different software

10:01APIs. Now, if you have no clue what APIs

10:03are, it's how automations are basically

10:05happening, how they're

10:07existent. I'm going to go through APIs

10:09in much more detail in the later stages

10:10of the video, so don't worry. But here,

10:12we're basically adding the username and

10:14password for each software. So, we

10:16always have it here. We don't have to

10:18re-upload it again. It's always a

10:19repository of passwords.

10:21Then we have executions, so we can see

10:23all the executions, which is, "Hey, this

10:25automation just run." It logs it, so you

10:27have everything logged and uh secured

10:29here, which is great. And lastly, it's

10:31data tables. Now, I'm going to dive into

10:33data tables in this module, I believe,

10:35or next module, um where I show you

10:36exactly how you can use data tables, how

10:38they're powerful, how they're better

10:40than using tools like Google Sheets and

10:42Airtable or Notion. So, we'll dive into

10:44that. Now, without further ado, let's

10:46press start from scratch, or you can

10:47press create a workflow. All right, and

10:49once you get to this dashboard, you'll

10:51be introduced with two different

10:52options. The first one is adding a first

10:53step, which is adding a first step

10:55manually right here, trigger, which is

10:57the first step of the actual workflow

10:59automation,

11:01or you can start using build with AI.

11:03Now, build with AI is only something

11:05that, depending on when you're watching

11:06this video, it only came out about a

11:08month ago, but when I press this button

11:10right here, it takes me to this right

11:12page right here. Now, the good thing

11:13about n8n is that it has two different

11:16types of support. It has the support

11:18where you can ask any questions about

11:20anything, and you can simply type any

11:22message here, and it will give you an

11:23answer, sort of like an n8n coach that

11:26helps you when you're building your AI

11:27agents or automations.

11:29And then we have the feature, which is

11:31build. Now, build again is something

11:33that recently came out, but it allows us

11:34to actually give it a prompt, so we can

11:36tell it, "Hey, what would you like to

11:37automate?"

11:38And it actually builds the thing. I'm

11:40going to just do this real quick, so you

11:42get to see exactly uh how it looks. I'm

11:44going to say, "AI agent that drafts

11:47my emails." And we have only 20 credits

11:49a month to use.

11:51I'm going to press go. As you can see

11:52now, this is thinking.

11:54It's searching through the nodes. I'll

11:56explain exactly what a node is.

11:57Essentially, it's a step in an

11:59automation. It's getting the node

12:01details, so it's understanding exactly

12:03automations in n8n, understanding which

12:05steps to put together and what. On the

12:07left-hand side, you get to see that this

12:08is working.

12:09Now, you get to see that it's adding the

12:10different steps of the automation. And

12:12the best way to describe this is that

12:13it's searching through a puzzle.

12:15It's understanding exactly what pieces

12:17do I need for the puzzle, and then it's

12:19getting the pieces, and then it's

12:21putting the pieces together right here,

12:23and then it's connecting, making

12:24connections inside those pieces, which

12:26we have right here.

12:28And now we can start using the

12:29automation. Now, I'm going to delete

12:30this and start from scratch, because I

12:32want to show you exactly what it looks

12:33like without the build with AI. As you

12:35can see now, we have no option to build

12:37with AI. We just have the first step.

12:38So, when I press this, I can see that on

12:40the right-hand side, it asks us, "What

12:42triggers this workflow?" Now, a trigger,

12:43like I mentioned before, is the first

12:45thing that starts the automation. Now,

12:46in n8n, we have the ability to trigger

12:48manually, so we can manually trigger

12:50this. We just press this button right

12:52here, which is execute workflow, which

12:53means that we are starting the

12:54automation, and it executes manually.

12:57We're able to trigger from an app event.

13:00So, something happens within any of

13:02these apps, we can get notified,

13:04and then it starts the automation. And

13:06there's about 500 apps here, a bit lower

13:08than other platforms, but n8n is

13:10constantly adding it here, which is

13:11great. We have on a schedule. This is in

13:13case you want to start your automations

13:15maybe once a day, or you want to run it

13:17once a week, or every other Tuesday on a

13:19July, you can do this. You can do on

13:21webhook call. Now, a webhook is

13:23something that I'm going to explain in

13:25the later stage of the video, but it's

13:27something that we use when we want to

13:28get notified when something happens.

13:30We have on form submission, which is for

13:32us to be able to use forms inside of

13:34n8n, so that when someone submits the

13:36form, it then sends us the information

13:38inside the form, and then we can start

13:40the automation. Then we can do when

13:42executed by another workflow, so we can

13:43actually have a workflow that calls

13:45another automation, right? So, workflow

13:47calls another workflow and then brings

13:48it data back.

13:50Then we have on chat message, in case we

13:52want to create a chatbot within n8n.

13:54Then we have when running evaluation, so

13:56this is in case you want to test the

13:58performance of your workflow or of your

13:59AI agent. And then you can do other ways

14:01as well, which in this case, it's using

14:04other different service. Now, I will

14:05cover this in the later stages of the

14:06video, but understand that a trigger is

14:09just the first step of the automation.

14:11Because let's say we do it manually, now

14:12we have the ability to then add the next

14:15steps. And so, when we have the first

14:16step here, the way that this works is

14:17that we have this connected to another

14:19step right here, and I can choose the

14:21next step. In this case, we have the

14:23ability to use AI in our workflows. We

14:25have the ability to use any

14:26applications, so any softwares, the

14:28ability to transform any data right

14:30here, the ability for us to have some

14:33sort of routing, which means that if we

14:35send the automation one way if this

14:36happens, the other way if something

14:38happens. And again, all these things

14:41might look complex because you're new to

14:42n8n, but trust me, they're much easier

14:44when you actually put them into

14:45practice. We have core, which is using

14:47code, HTTP, webhooks, a bunch of stuff

14:50that you shouldn't worry about now, and

14:52human in the loop, which is another

14:53feature that I'll be explaining. So, we

14:55have the first step, then we have the

14:56next step. So, let's say I want to talk

14:59to Notion

15:00or Google Sheets. Let's say Google

15:01Sheets.

15:03Now, we are introduced with Google

15:04Sheets, and these are all the actions

15:05that we can do within that software,

15:08right? So, with every software that we

15:09have, let's say it's even Gmail. If I

15:11press on the actual software, I can see

15:13that all the actions that I can do

15:14within that software. I can add a label

15:17to message, delete a message, get a

15:18message, get many messages. I can do a

15:20bunch of stuff within each software,

15:22which is where automation truly shines,

15:24because it does the things that we don't

15:25want to do, or that we don't have to do.

15:27And so, let's say I wanted to go back to

15:29Google Sheets, I can just press Google

15:30Sheets here, action in app, which you

15:32find again right here, action in app.

15:36So, then

15:36I can go Google Sheets.

15:38I can choose the action that we want to

15:39take, which is append a row.

15:42And now we're introduced to this page

15:43right here. Now, every n8n node has an

15:46input,

15:47then has in the middle the

15:48configuration, which is, "How do we

15:50actually set the thing up?" And then the

15:52output, which is, "Okay, using the

15:54configuration here, what was the

15:55output?" All right, so I just made a

15:56Google Sheet here that has name, email,

15:58phone number, company, role, and

16:00address. Let me name this contacts. Now,

16:02here, if I go back to n8n, the first

16:04step to building any kind of automation

16:06with different steps in the flow, is

16:08connecting our accounts to these

16:10softwares. So, we're able to access them

16:13without us having to go to the actual

16:14software to do it. So, right here, you

16:15get to see on the top that we have

16:17credential to connect with, which is,

16:19just press here, you can create a new

16:20credential, which means that you're

16:21connecting your account, and you're

16:23introduced to this page. Now, if you see

16:25OAuth 2, service accounts, allowed HTTP,

16:27and you get overwhelmed, don't worry,

16:28just sign in with Google. It's just a

16:30simple login verification, and you're

16:32all good. I'm going to press the account

16:34that I want to connect to. You can press

16:36select all, so you're basically saying,

16:37"Hey, I trust n8n to have access to my

16:41Google Drive and my Google Sheets." You

16:42will get to this page, which is

16:43connection successful, and you'll be

16:44redirected here to n8n. I can make a

16:47name, so I can say Michele

16:49system

16:53account.

16:54So, I'm pressing save. And usually, what

16:56we have is also the date. And the reason

16:57why we add the name is because when you

17:00connect 20 to 30 accounts here, you want

17:03to know exactly what the account was.

17:05And so, the only way for us to know

17:06exactly what the connection is, is to

17:08name the actual connection. Once you

17:10have this here, now the resource is

17:11usually, "What is the thing that we're

17:13manipulating? What is the thing that

17:14we're taking and doing something with?"

17:16In this case, it will be a Google Sheet.

17:17The operation is always the action that

17:19we're taking. So, as you can see here,

17:21we have different options. We have

17:22create, delete, delete rows, get rows,

17:25update a row. You can do a bunch of

17:26stuff. And in this case, you have append

17:28a row, which is create a new row in a

17:29sheet,

17:30right here. The document, it's choosing

17:33the actual Google Sheet, which in this

17:34case is contacts, right here. See how I

17:37just right here. And also, what you

17:39could do is actually go to the ID

17:42or the URL

17:44of the Google Sheet, which I can find,

17:45if I go here, I can find in the URL

17:46here, and the ID is usually

17:49right here. And then the sheet, which is

17:51the sheet within the Google Sheet, which

17:53in this case is sheet one. That's a lot

17:54of sheet right there. Um and finally,

17:56you have this right here. And so, this

17:58is where you start to map columns. So

18:00again, what we're doing here is we're

18:01adding a row inside here. So, we're

18:03adding a row into name, email, phone

18:05number, company name, role, and address,

18:07which is why when I go back here, I can

18:08see that it's asking us, "Hey, what are

18:10the different values that you want to

18:11send?" In this case, for name, let's

18:13just put James Arthur. Let's do email,

18:16my email here, my phone number,

18:18my company, the role,

18:21and the address, so on. And now, leave

18:24everything else as is. These are just

18:26other options that you can um sort of

18:28play around with, but there's no need

18:29for that. And all we have to do now to

18:31actually run the automation is to be

18:35able to connect this

18:37to here.

18:38So, this is saying, "Hey, the first step

18:39now is connected to the next step." And

18:41so, if I press here execute workflow,

18:43this should now start the automation and

18:46do the action, append the row, like we

18:48mentioned, and adding these details in a

18:50new row.

18:51So, let's try it out. Let's say I go to

18:53execute workflow. I can see that the

18:54first step executed, and we know this

18:56because it was all green, check mark,

18:58and the second step as well. And if I go

19:00to my Google Sheet, I can see now that

19:02James Arthur, with his email, phone

19:04number, James Solutions, CEO, and Spain,

19:06was added to our database. And that

19:08right there was a full execution from

19:09the start until the end, workflow

19:11executions, from the first step until

19:13the last step. Now, we can add another

19:15step, which let's say is Gmail. I chose

19:17to type Gmail. I can choose the software

19:19here, and now I can choose the action.

19:21So, in this case, it can be send a

19:22message.

19:24And now we're introduced to this page

19:25right here, which is very, very similar

19:27to the one that we had on Google Sheets,

19:28but the only difference now is that we

19:30have access to the previous, we call

19:32them variables, right? Because they're

19:33different variables here.

19:35And now we can use these variables in

19:37the automation. And that's why we don't

19:38call them variables, what we call them

19:40dynamic variables, because they change

19:42every single time. Because when I put

19:43name somewhere here,

19:45and I run the automation again later, if

19:47the name changes, this automatically

19:49changes name. So, all I have to do here

19:51is do the same thing. So, I can send a

19:52message, which in this case is send an

19:54email. I can sign in with Google. I can

19:56choose my account, select all.

19:58Connection successful.

20:00And now [snorts] I can name this, press

20:01save. Correct. And now we've connected

20:03our account to n8n. So, now n8n is able

20:06to use our Gmail. The resource, which is

20:09what is the thing that we're

20:10manipulating, again, changing, is a

20:12message, which in this case is an email.

20:14And you can use draft, thread, label,

20:15you can do whatever.

20:16Operation, which is what is the action

20:18that we're taking? In this case, it's

20:19sending. We can do a bunch more stuff

20:21here, depending on what you want to do.

20:23And then it's asking us, who do we want

20:25to send the email to? What is the

20:26subject line of the email? And what is

20:28the message of the email, which is the

20:29email body. So, now, as you can see

20:31here, when I pressed on two, it gave me

20:33the ability to drag an input field from

20:35the left to use it here. So, now I can

20:37take the field from the previous step,

20:39drag this across over here,

20:42email, and have this be the variable

20:45that we use every single time, instead

20:47of putting the email here, right?

20:49Instead of putting the email manually.

20:51Get it?

20:52Get it? 25@gmail.com.

20:55Because if you do it this way, it's

20:56going to be fixed. But, if you take the

20:58variable from the previous step, and I

21:00drag this across here, then only are we

21:02able to change the variable dynamically,

21:04without us having to go here and change

21:06it manually ourselves. And so, what

21:08we're doing now is that we've just

21:08pulled in the variable here. We can do

21:10name. So, pull this name here,

21:13and say, "Welcome to the team." And

21:16also, one more thing that you start to

21:17see is JSON code, right? Now, this looks

21:20very intimidating for someone looking to

21:22get started with automations and going

21:24into n8n, but trust me, it's actually

21:26quite easy. Because this code right here

21:28is literally just the variable here that

21:30we just pulled across.

21:31So, you don't really have to know

21:32exactly what this is. But, in case

21:34you're wondering, we can actually

21:36hardcode it. So, we can do curly

21:38bracket, curly bracket. This opens this

21:39up. You can do JSON {dot} the actual

21:43step from the previous step. So, you can

21:45go to, let's say, email. And you when

21:47you press email, it automatically takes

21:49us back to the same place as if I were

21:51just to drag this across.

21:53And this is great for us in case we want

21:55to reference any variables from the

21:57previous steps, and there's a ton of

21:58variables, and we just want to reference

22:00them really, really quickly, we just

22:01manually hardcode it. But, most of

22:03times, you're just going to drag and

22:04drop. And then, we're going to ask the

22:06email type. So, in this case, let's just

22:07do text. HTML is just the way that you

22:10make your emails look pretty. We can

22:11say,

22:12"Hey brother,

22:14happy to work with" And let's do

22:17company. Company right here.

22:19So, now we're saying,

22:21"Hey brother, [snorts] happy to work

22:22with JSON variable, which is company,

22:24which is this name right here."

22:26We just don't see James Solutions

22:27because

22:29it's a variable right here. Now, the

22:30good thing about n8n is that we can

22:31actually see the output.

22:33So, we can see the actual variable

22:35before we run it, which is amazing,

22:36because now it this gives us the ability

22:37to visualize whether it's correct or

22:39not. So, James Arthur, welcome to the

22:41team. "Hey brother, happy to work with

22:43James Solutions." Which is great. Now,

22:45what I can do is go back.

22:47I can go back here. I can change the

22:48email to, let's say,

22:52this one here that I have,

22:54migdallahcalde5@gmail.com.

22:56And now, all I have to do is execute the

22:58workflow again from the start until the

22:59end.

23:01Right here.

23:03This now goes into Google Sheets, and it

23:04goes here. I can see now that I have the

23:06name, the email, and I guess the email

23:08changed, right? Because we changed it

23:10manually in n8n. The phone number, James

23:12Solutions, CEO in Spain. And as I showed

23:14you, it also sent the email. So, if I go

23:16to my email here, I can see that we had

23:18a new email called James Arthur,

23:19"Welcome to the team. Hey brother, happy

23:21to work with James Solutions." Which is

23:23the exact same email that we added into

23:25our n8n. And so, now that we've built

23:27this, I can explain exactly how the

23:28pricing works. So, from the start until

23:31the end,

23:32right? When you are testing, so when

23:34we're pressing this button and we're

23:36executing this,

23:37we're not paying for it. From the time

23:39that we want to set this live, so we

23:41don't have to manually do it every

23:42single time, we can turn this on by

23:44pressing this button.

23:45So, right now, the workflow, or the

23:47actual automation, is inactive, which

23:49means that the automation is only

23:51running when we manually run it. When we

23:53go inside the platform, when we manually

23:55press this button, and it goes here.

23:57But, in the case for clients that you

23:58want to set this live, it's sort of like

24:00this is just for testing. When you're

24:01building automations, you're testing

24:03here and there. And the live is when you

24:04want to actually set this live, it goes

24:06live. Now, it works by itself, and it

24:08runs through this at whatever interval

24:10that you want. And the good thing about

24:12this is that we're not paying for it

24:13when we're testing, which is great,

24:14because it allows us to be able to spam

24:16the testing as much as we we can,

24:18because the way that I they actually

24:19charge you is by workflow executions,

24:21which is from here to here. As you can

24:23see, we have the green check marks,

24:25which tell us that everything went good,

24:26which is awesome. Now, let's say

24:28something went wrong, right? So, if I go

24:30here, and I just create a new

24:31credential, and I just save it. So, I

24:35just put a fake credential. Close.

24:38Gmail account. This shouldn't work. So,

24:40now, if I execute the workflow, I can

24:42also execute it here, instead of going

24:44here, I can see that now in the bottom,

24:46you can't see it because my face is

24:47there, but if I just move myself here, I

24:50can see that I have problem in node

24:52here.

24:53Right? Which allows us to see that we

24:55have a problem in our automation. And

24:57so, this node right here, again, a node

24:59is just a square within n8n, is a step

25:01in our automation. We can see that we

25:03have an issue, and it tells us the

25:04actual issue, "Unable to sign in without

25:06access token." Which means that the

25:08authorization, the actual account, was

25:10invalid. And on the right-hand side, you

25:12can also see if something went wrong,

25:14what is it that went wrong. And you can

25:16also ask AI, which is amazing, because

25:19this AI is actually good. It's not some

25:20sloppy thing that they just added into

25:22the software. It actually works, and

25:24it's really good. With that said, make

25:25sure that when you build automations,

25:26you always save right here on the top.

25:28And up here, you have the ability to

25:30edit. So, this is where we edit our

25:32workflows. Executions is where we get to

25:34see all the times that this executed,

25:37that this ran, right? The first time it

25:38ran, the second time when we had more

25:41steps, the third time when we had even

25:43more steps, right? So, we get to see

25:45exactly the previous times that it ran.

25:47And then, we have evaluations in case

25:49you're running some more advanced

25:51evaluations, you're testing your your

25:52actual workflow. If you're building this

25:54for clients, you're most likely going to

25:56use this, but it is not necessary to

25:58start. Going back to editor right here,

26:00on the top right here, I have the

26:01ability for me to duplicate the actual

26:04workflow. So, if I duplicate this, I

26:06have to rename this, and I can press

26:07duplicate. And this will now make the

26:09exact same workflow here. Then, I have

26:11the ability to download. So, if I

26:13download it now, I get a JSON file. And

26:15so, the good thing about that is that

26:16this, in the back end, is all JSON,

26:18which is is all code. And so, I can

26:20simply just press download, like I

26:21showed you, and then you can press

26:23import from file, which allows you to

26:25import the automation into your own

26:26workspace. So, if I go to a blank

26:28workflow, I can simply press this button

26:30right here. I can import from file.

26:33And now, I can put the actual file that

26:35I need.

26:36And suddenly, I have the actual

26:37automation, which is imported into your

26:39my n8n account. And then, finally, we

26:40have change the owner, so you can change

26:42the owner of the actual automation. You

26:44can rename the automation this right

26:46here. And you can import from URL. So,

26:48in case you turn the file into a URL,

26:50you can simply import it there. On the

26:52right-hand side here, like we mentioned,

26:53we have the ability for us to add steps.

26:56We have the research, so we can command

26:58bar, so you can actually search through

26:59the different nodes that we have here.

27:01We have notes, which is a big part of

27:03n8n, because we can actually start

27:04documenting our automations. So, I can

27:06go in here,

27:07and I can say, "First step is

27:10manual."

27:12And I can delete all of this, and I can

27:13put this over the automation, which is

27:15great, because when I'm going to give

27:16the automation to someone, right? I can

27:18actually give them an explanation of

27:19exactly how it's working. If I go here,

27:21I can also say, "This

27:23step allows us to be able to manually

27:28trigger the automation."

27:31Right? And so, we have the text here.

27:32And you can start to see how bigger

27:34automations and bigger workflows use

27:36this. And it's something that I use all

27:38the time when I have to give workflows

27:40and AI agents to my community members

27:42who are looking to get them and then

27:43start using them right away. And now

27:45they have an explanation of what it is.

27:46Then, we have right here focus panel, in

27:48case you're doing some note stuff here.

27:50We don't really use this that much, so

27:51don't worry. Um and finally, we have the

27:53AI, which I mentioned is able

28:02or is able to assist us whenever we have

28:04any questions about automations. Now,

28:06you probably might be asking yourself

28:07this question, right? Which is, if you

28:10can build with AI, why are we learning

28:12this at all? It's because AI only gets

28:14you so far. When something doesn't work,

28:17how do you know if it doesn't work? How

28:18do you know if it's not good enough? How

28:20do you know if it's not efficient? Which

28:21is where understanding the fundamentals

28:23is very, very important, rather than

28:25relying on AI for everything. All right.

28:26So, now that you understand exactly what

28:28automation is, how you can set up your

28:30own n8n account, how n8n actually looks

28:33visually, plus building your first

28:34automation inside of n8n, now it's time

28:37to go a level deeper. Because in the

28:39next module, we'll dive into actually

28:40understanding how automations really

28:42work. We'll dive into APIs, webhooks,

28:45and how data flows within automations.

28:47We will look at credentials. We will

28:49look at transforming data, and also

28:51understanding error handling, so if

28:52something goes wrong, what do we do

28:54about it? And lastly, understanding the

28:57core 17 nodes that you need to know when

28:59you want to master n8n. So, by the end

29:01of this section, you will not only use

29:03n8n, but you will actually understand

29:05it.

29:12If I had to learn just one skill for AI

29:14automation, it would be understanding

29:16how APIs work, because fundamentally,

29:18they are the things that allow softwares

29:20to speak with other softwares. And so,

29:22once you understand, really understand

29:23how APIs work, you'll be able to use

29:26n8n, Zapier, make.com, Power Automate,

29:28all with no problems, without getting

29:30stuck, and without any limitations. And

29:32the problem is, most people skip this.

29:34They go straight to the platforms to

29:35build automations step by step, and when

29:37something breaks, they have no clue why.

29:39So, in this video, I'm going to break

29:41down exactly how APIs actually work, how

29:43to read an API documentation without

29:46freaking out, and how to build your

29:47first real API request inside of n8n.

29:50Let's dive in. All right, so API stands

29:52for application programming interface.

29:55>> [music]

29:55>> A very fancy term that nobody even knows

29:57the meaning to. But what it means is the

30:00ability for softwares to speak with

30:02other softwares. Now, think of a waiter

30:03in a restaurant, right?

30:05This waiter has the chef, and he has the

30:07customers. And between the chef and the

30:09customers, there's a series of actions

30:10that happen, right? So, take the order,

30:12bring the order from the kitchen, they

30:13deliver the order, make the order.

30:15There's a whole repetitive loop that

30:16happens over and over again. And so,

30:18that is the exact same when it comes to

30:20APIs for softwares. Where we have one

30:23software here, one software here, and

30:25these two softwares communicate to each

30:27other through a series of requests and

30:29responses. So, request in this case

30:31would be the make the order, the

30:32response would be the actual order being

30:34delivered to their table. And that is

30:35the exact same when it comes to the

30:37softwares. And so, fundamentally, the

30:39waiter in this case is the API, and the

30:41chef is one software, and the customers

30:43are other softwares. And this right here

30:44is the thing that allows the softwares

30:46in make.com n8n to speak to each other,

30:48to be able to have a sequential order.

30:50So, we send data to a place and get

30:53something back, and that's how

30:54automations and AI agents are built.

30:56Now, what can APIs do with softwares?

30:59There's a few things that APIs can do.

31:01We call them requests. So, requests are

31:03just a way for the server, the API

31:05server, to send information to the

31:08actual software itself to then create an

31:10action. Now, the first thing it can do

31:11is post.

31:12So, this is when any type of new record

31:14is created in a software. So, this is

31:16like making a Google Sheet row. It's

31:18creating a contact on ClickUp, a contact

31:20on HubSpot. It's

31:22you know, sending information to OpenAI

31:24to make a LinkedIn post. This is a post

31:26request, which is sending information

31:27from one place to another. Then we have

31:30get. So, get is extracting any details

31:33from any record from the software. So,

31:35let's say someone filled out a form on

31:37your website, you wanted to get the

31:39information from that form, and then use

31:41it for whatever you want. Right now,

31:43you're getting information because the

31:44information is going from somewhere else

31:46to us. That's getting information.

31:49Then we have put or patch, which is

31:50updating records in a software. So,

31:52let's say John's name is not John

31:54anymore, it's James, then you use a put

31:56or patch request, right? So, updating

31:58information. And then we have deleting.

32:00So, you can't see this fully, but it's

32:01deleting a record from a software. So,

32:03let's say we had a Google Sheet, or

32:05maybe a contact, we just want to delete

32:07the row, or delete the Google Sheet, or

32:09we just wanted to delete the contact.

32:10That is a delete request. So again, post

32:13is to send information, get is to get

32:15information, then put or patch is to

32:16update information, and then delete is

32:18to delete information from somewhere.

32:20Now, I'll share a story here. So, I was

32:21on a call with a client a while back,

32:23and she asked to automate different

32:25parts of her onboarding process. And she

32:27used Zoho CRM as the CRM, which by the

32:30way is just a place for companies to

32:32store information, to be able to send

32:34information to Zoho to make a new

32:36contact whenever someone filled out a

32:37form. Now, the problem is I had never

32:40used Zoho CRM. So, I'd never used it. I

32:42didn't know exactly what you can and

32:44cannot automate with Zoho CRM, which is

32:46why we get to the next part, which is

32:47finding the information of what you can

32:49automate and cannot automate within a

32:51software. So, what I did is a simple

32:53Google search, and I searched up the

32:55software name.

32:56So, in this case, it can be Zoho CRM,

32:58Monday, it can be OpenAI, it can be

33:00whatever it is.

33:02API documentation. And this is where we

33:04get to the fundamental concept of an API

33:07documentation, which allows us to be

33:09able to see exactly what you can

33:11automate within a software and how to

33:13set everything up. Now, here's an

33:15example of an API documentation. This is

33:18Apollo's API documentation, and at

33:20first, you think this is way too

33:22complex, but the reality is, if you

33:24understand how this looks, all the

33:27others will be very, very easy because

33:29they all follow a very similar

33:30structure. So, the way that an API

33:32documentation works is that on the

33:33left-hand side, right here, you get to

33:35see all the types of requests. So, this

33:37is enriching people, it's searching

33:39people, searching organizations,

33:40creating an account. There's different

33:42things that you can do within Apollo's

33:43API, which means automatically without

33:46having to go to Apollo, right? And so,

33:48each action right here, so each thing

33:50that we can do has a corresponding type

33:53of request that follows through. [music]

33:54So, people enrichment, in this case,

33:56would be a post request because we're

33:58sending information to the server, and

34:00then it's enriching the person. Then we

34:02have, let's say, organization

34:03enrichment, which is a get request, or

34:05update an account, which is a put

34:06request, right? So, we have different

34:08types of request and different actions

34:10that we can do using Apollo

34:12automatically within our automations.

34:14And the only way to know is actually

34:15just looking here. Now, let's say we go

34:18to people enrichment. So, this will now

34:20look at what you can do within that

34:22specific request, which in this case is

34:24enriching people. And this request right

34:26here has the type of request, and then

34:28has a URL. Now, the URL is something

34:31that is very, very important in our API

34:33requests because they are the thing that

34:35allows us to be able to send the

34:37information the right way. It's sort of

34:39like going to google.com, right? The

34:41server knows that you want to go to

34:42google.com because

34:44because of the request, because of the

34:46URL. And so, this is the exact same

34:48because this allows the server to know,

34:49"Hey, he wants to use Apollo, and he

34:51wants to enrich people." Then in the

34:53middle here, it gives us a summary as to

34:55what it is, how it works, and so on. So,

34:57these down here are basically saying,

34:58"Hey, this is the type of request." So,

35:00right here,

35:01this is the information you have to send

35:03to the server to call the request, to

35:04make it happen. In this case, it's

35:06enriching people, and you have to send

35:08the first name and the last name, right?

35:10And those are the details that you have

35:11to look into understanding exactly what

35:14to send to the server. And on the

35:16right-hand side, you get to see the the

35:19actual setup. So, how can we set this up

35:21in our automation platforms? And quick

35:23side note, the reason why we look at the

35:25API documentation is because here, if I

35:27go to Apollo,

35:29I can't find it. Which means that now we

35:31have to set up the API request using the

35:33documentation. And so, the left-hand

35:35side is the type of request, the middle

35:37is to show you the exact request and

35:39sort of to give you a summary of what it

35:40is and what it looks like. And the

35:42right-hand side is saying, "Okay, we

35:43have the request, we have the summary,

35:45here's how you can set it up." Now, the

35:47main thing that we have to look at when

35:48we have to set up an API request through

35:51an API documentation is the curl. Now,

35:54the curl is the thing right here that

35:55you see,

35:56which basically gives us a heads-up or

35:58an understanding of how the request

36:01actually works. And if I go to n8n right

36:03here, and I go to HTTP request, which is

36:05the thing that we use to make the actual

36:07request, we can see that we have a

36:09button right here, which is import curl.

36:12And this will be the place where we

36:13import the curl, which is the one that

36:15we saw on the API documentation, and

36:17then this basically makes the whole

36:18request for us, which I'll show you in

36:20just a second. Now, this will make much

36:21more sense once we actually build the

36:23API request, which is super damn easy,

36:25so don't worry. Uh we'll go through it

36:27step by step. But now you get to see

36:29what a structure of an API documentation

36:31looks like. And this is for a software

36:33called Apollo. Now, if I go to the

36:34second API documentation, this right

36:37here is for a software called Instantly,

36:39which is a cold email software. As you

36:41can see, it follows the exact same

36:43structure. On the left-hand side, we get

36:44to see all the actions that we can take.

36:46On the middle, it tells us this is the

36:48summary, or this is what it is, and

36:50these are the different variables that

36:51we have to send to the server in order

36:53to make the action happen. And on the

36:55right-hand side, this is the way that we

36:57set it up through a curl, which is in

36:59code, right? But again, we don't have to

37:01know how to code. It's just telling us

37:03the way that we have to set things up.

37:05And usually, the curl in this case has a

37:06URL, which is the action that we're

37:08taking, telling the server, "Hey, we

37:10want to enrich leads, we want to send

37:12this information somewhere else." And

37:13then we have the headers, which usually

37:15are a way for the actual server to

37:17authorize that this is your account. And

37:19the way it does that is through

37:20something called an API key, which we

37:22use for tons of things. OpenAI API key,

37:25we use Claude API keys, Perplexity API

37:27keys. It's just sort of a way for the

37:29server to know, like a password, to say,

37:31"Hey, this is a key, it's only for me,

37:33for my account, so you can take action

37:35on my account, and you know it's my

37:36account if it's this key." And then we

37:38have content type application.json.

37:41Sometimes it's there, sometimes it's

37:42not. It's just a way for us to tell it

37:44to format the information in the right

37:45way.

37:46And then D, which is the body, which are

37:48these parameters, so these variables,

37:50which we send through to the actual

37:53server to make things happen. And if

37:55that confused you even more, then just

37:57wait until we build the API request cuz

37:59it'll be very, very easy. And then the

38:01last API documentation will be ClickUp.

38:03As you can see, I'm just showing you the

38:04example, so you get to see the

38:06reoccurrence in the structure of an API

38:08documentation, which is the one right

38:09here. On the left-hand side, we have the

38:11types of request. In the middle, we have

38:13the summary, the URL, the type of

38:15request, which is post, the title, and

38:17then here will be the parameters that we

38:19have to send through to the server to

38:22say, "Hey, here's what you have to do,

38:24here's how to do it," and so on. And on

38:26the right-hand side, this is the way

38:28that we set it up, which is the exact

38:29same. And as you can see, the thing that

38:31we really care about is the curl request

38:33because this is the right-hand side part

38:35of these API documentations, which allow

38:38us to be able to know exactly how we set

38:40things up. And if you read API

38:41documentations a lot of the times, or

38:43most of the times, you'll see something

38:45that looks like these. These are three

38:47random API documentations, and they all

38:49follow the exact structure. So, let's

38:51set up the API request. Let me give you

38:54a use case. Let's say we want to get the

38:55weather of a country or a location,

38:58right? If you go here to weather, we can

39:00do the OpenWeatherMap, but there's

39:02actually this other software, which is

39:04called the Weather API. So, if I go here

39:06to Weather API,

39:08there's this software right here, free

39:09weather API,

39:11which isn't listed here, right? And this

39:13is when we start using the HTTP request.

39:16Okay? So, HTTP request, again, like I

39:18mentioned before, is all the case where

39:22we can't find the exact software

39:24for the specific action that we're

39:26taking. So, we actually make our sort of

39:28custom API call, which is saying, "Hey,

39:31even though we might not have it in n8n,

39:33it doesn't mean it's impossible, right?

39:34We can still make the actual request,"

39:36which is where the HTTP request comes

39:38in. So, all I have to do is go here, put

39:41HTTP,

39:42and now we get to this page. The first

39:44thing is the type of request, whether

39:46it's a get, delete, head, options, path,

39:49post, put, which we went through just

39:50now. And we have options head, which

39:53I've never used, so probably not

39:54something that you should know about.

39:56Then we have the URL, which is the

39:58action that we're taking, which we get

40:00from the API documentation. Then we have

40:02the authentication, which we'll go

40:03through in just a second. And then we

40:05have the parameters, which I mentioned

40:06will be here. And then also the body,

40:08which is

40:09this right here, right? Or you can also

40:11use JSON, which we get from the API

40:13documentation. Now, we'll have to set

40:15one up so you understand. And let's go

40:17to the free weather api.com. We get

40:19here, and now I'm saying, "Hey, I don't

40:20have it within any 10. So, I really have

40:22to go to the docs." And usually you find

40:24this here.

40:25So, docs. All right, so this right here

40:27is the API documentation.

40:29And so, all we have to go is to the

40:31request, right? So, we have the getting

40:33started, we have the authentication, we

40:35have the request URL, we have a bunch of

40:37things, but the thing that we really

40:38care about is the bulk request, which is

40:41the request, which will be up down here,

40:44I believe. Yeah, down here. Which will

40:45be the request that we will use as an

40:48example

40:49for the actual API request that we do

40:52within any 10. So, all we have to do is

40:54copy this, go here, and set up an HTTP

40:57request, import curl, just paste it

41:00here. Now, we can see that everything

41:02was done for us. We have the post, we

41:03have the URL, we have the fields, the

41:05key, the queue, which in this case would

41:08be um

41:10We can see here, we can see that we can

41:12do the city name, we can do the US zip

41:14code, UK post code, and so on. Um

41:17which is required. And then we have the

41:19API key, which is the one right here,

41:21which is sort of a password that says,

41:22"Hey, this is my account. This is um

41:25related to me. Just use mine." And the

41:28body in this case, we actually don't

41:29need this because this is just an

41:30example.

41:32So, I can delete this. And we can just

41:34work with this. So, type of request,

41:35which is post, URL, which is the URL

41:37that we use. So, what is the action that

41:39we're taking? And then we have the key,

41:40which is the API key, which we find on

41:42my account. So, when you make a free

41:44account, you can go here

41:46to the dashboard, and up here

41:48you will copy the API key

41:50and paste it back here.

41:52And then the queue will be the city

41:54name. So, we can do London

41:56or whatever it is.

41:57And I press execute step, which will now

41:59execute the step. So, as you can see

42:01here, it successfully ran this. And I

42:03get the weather,

42:05so the current weather, the temperature

42:07of that specific place,

42:09which you can then use for whatever it

42:10is. Right, so now you got to see exactly

42:12what it's like to build an API request,

42:15an API call, which we [music] call. And

42:17this is exactly what you can then use in

42:19between your automation steps. So, let's

42:21say we had a next node, which would be

42:23let's do an edit fields, and we wanted

42:25to get the temperature.

42:28So, temp, temperature

42:31in Celsius, 9.3. Yeah.

42:34Thank god I'm not in London.

42:36I can execute the step, and this will

42:38now call the server, bring the data

42:40back, and then here you can see that the

42:41temperature was 9.3 because this is a

42:43variable that we found here. But again,

42:46the reason why we're using HTTP is

42:48because if I go here to free weather

42:49API, it's not listed here. It's not

42:51here, right? And any 10 only has about 4

42:54to 500. It has the main softwares, but

42:56sometimes softwares that we want here

42:58are not here, which is exactly why we

43:00have to then go to the API documentation

43:03and find or understand how we can set up

43:04the API request.

43:09In this video, I'm going to show you how

43:10you can master any 10 just by learning

43:1217 nodes. And these are the exact same

43:15ones that we use pretty much 80% of the

43:17time when building automations for

43:19clients. And honestly, these are

43:20probably the only ones that you'll ever

43:22need. So, the first one is the trigger.

43:24Now, the trigger is essentially the

43:25first step of the automation. Like, what

43:27is that thing that starts the automation

43:28for us to then do the next steps? In any

43:3010, we have different types of triggers.

43:32The first one is a manual trigger, so we

43:34can just execute the workflow manually.

43:36Now, this right here is something that

43:37we use most of the times just for

43:39testing. [music]

43:40Right? Because you want to test

43:41something, so there has to be a first

43:43step, and that's where we add a manual

43:45trigger. Now, the second one is a

43:47schedule trigger. So, this trigger right

43:48here is primarily used when you want to

43:51run the automation, let's say once a

43:52week, once a month, or at certain

43:54interval of time. So, if I go in here, I

43:56can see that I can run it

43:58in seconds, minutes, hours, days, weeks,

43:59months, and even custom. So, this is

44:01cron, which you can put your maybe every

44:032:00 p.m. on a Tuesday or every 3:00

44:05p.m. on a Friday of every month. And you

44:07have different settings that you can put

44:09so that you're able to

44:11to run the automation at a very specific

44:13time of the month, of the year, of the

44:14week, of the day, right? So, to give you

44:16a use case, let's say we had a content

44:18system. Let's say we wanted to generate

44:20content automatically every single day.

44:22What we would do here is we would go to

44:23trigger interval, which is days,

44:25one, midnight, [music] and trigger

44:27minutes zero. And this is where we put

44:29this as the first step of the automation

44:31so that it runs every one day at

44:33midnight of the week, right? So, we're

44:34able to run the automation without us

44:36having to manually trigger it. And this

44:38is for the time trigger. And then we

44:40have the Typeform trigger. Now, this is

44:42just to show you an example of a trigger

44:44that is not native to any 10. These two

44:46are native to any 10, which means that

44:48any 10 owns those nodes. But in this

44:50case, Typeform is not any 10, it's a

44:52different platform. And so, Typeform

44:53itself is a platform that generates

44:55forms. So, we can generate forms, and

44:57this trigger right here starts whenever

44:59someone fills out the hiring form in

45:01this case. So, I have the form pulled

45:02out here, which is a form that is

45:03connected to this Typeform trigger. And

45:05what happens is that whenever someone

45:06fills out the form, this sends a signal

45:09to this node because it's an on app

45:11trigger

45:12that will allow us to then be notified

45:14and start the automation. So, I'm going

45:15to press execute workflow. As you can

45:17see now, this is waiting for you to

45:18create an event in Typeform, which is

45:20waiting for me to fill out the form. I'm

45:21going to press submit. And as you can

45:23see here, this is now triggered. And we

45:25got the data, the name, the email, the

45:27phone number, and location as well.

45:29Right? So, this essentially is an on app

45:30trigger, which means that it triggers

45:32whenever an single app starts something

45:34or whenever something happens within an

45:36app that we can all find

45:38right here. Add another trigger, we have

45:40all these triggers, and the on app event

45:43is the one that you can use for whatever

45:45it is. So, these are all different

45:46softwares that can be the first step of

45:48the automation that triggers whenever

45:49something happens within that software.

45:51Now, the next part of the nodes are

45:53going to be storage solutions. Now,

45:54storage solutions is just a fancy way to

45:56say

45:57we use nodes that allow us to store

45:59information in them, right? And so,

46:01typically we use Google Sheets, we can

46:03use Airtable, we can use Notion, we can

46:05use any 10's new native data tables,

46:07which only came out a few days ago, in

46:08order to store information. And this

46:10could be a simple Google Sheet that

46:11looks like this, which has full name,

46:12email address, location, and phone. But

46:14it could also mean a data table within

46:16any 10, which has name, email, phone

46:18number, and location, which is the exact

46:19same. The only difference is that one is

46:22stored in any 10, and the other one is

46:24stored in Google Sheets. So, let me show

46:25you exactly what I mean here. Let me put

46:27the

46:28um manual

46:29execution here.

46:31And I can see that inside, I am adding a

46:34row. So, append row means I'm adding a

46:35row in the any 10 17 nodes, which is the

46:38name of the Google Sheet right here.

46:40And I'm putting the first name or the

46:42full name as Michele Torti, the email as

46:44this, and the location as this. I can

46:45press execute workflow, which will now

46:47run the step right here.

46:49And it will add the row

46:51into the Google Sheet, which will be

46:52Michele, email, and location as well.

46:55And I can also add the phone, which you

46:57can also add by refreshing.

47:00Let me add let's do 1 555 123 4567.

47:05And if I execute the step again,

47:07I can now see that I will have a new row

47:09that will be added here with email,

47:11location, and this as well. Now, the

47:12reason why it gave us an error is

47:14because Google Sheet doesn't like when

47:15you put plus signs or equal signs and so

47:17on because these are native equations

47:20within Google Sheets.

47:22But you get the gist here. You just

47:23store information into Google Sheet. The

47:25second one is within the any 10 data

47:26tables. So, let me connect this. And in

47:29this case, we do the exact same thing.

47:31It will be new people because that is

47:33the name of the database.

47:34And we're adding the first name or the

47:36name as Michele Torti, my email, my

47:38phone number, and location as well.

47:40And if I press execute step,

47:42I can see that now it was successful.

47:44And if I go here and I refresh, I can

47:47now see that the data was added here.

47:49Full name, email, phone number, and

47:51location as well. Now, the amazing thing

47:53is that once you master Google Sheets

47:55and any 10's native data tables, you can

47:57pretty much use Airtable, you can use

47:58Notion, you can use Asana, ClickUp. All

48:01these places where you store

48:02information, they're all very, very

48:04similar, right? So, if you understand

48:05one, the setup will be very, very

48:07similar for the other ones. All right,

48:08so now we get to the universal data

48:10processing. So, here is where we

48:12manipulate data, right? Manipulate data

48:14just means that we take some sort of

48:15data that comes in, and then we

48:17structure it in a way where it makes

48:18sense for us to structure it in. Now, in

48:20this case, I want to take the um manual

48:23execution here.

48:24I want to attach it here.

48:27And I can see that inside the edit node,

48:29if I execute the step, I can see that I

48:31have an array, which is basically a list

48:33of people's information. So, we have

48:35name, email, phone number, and location.

48:37David Smith, David Smith, Emily Chen,

48:39Carlos Ramirez, and Aisha Khan. And

48:41these are basically giving me the

48:43different pieces of information. Now,

48:44the only problem with this

48:46is that let's say we want to add this to

48:48our Google Sheet,

48:49we can't because this is an array, which

48:51means that it's inside

48:53this right here.

48:54So, this is where the split out comes

48:56in.

48:58So, what this means is that if you want

48:59to process each name individually,

49:01right? If you want to process this right

49:03here individually, then this right here,

49:05then this right here, and this right

49:06here, and this right here as well,

49:08then we have to split them out. We have

49:09to take them, as you can see by the

49:11diagram here, it's taking them, which is

49:12a singular thing, and then it's

49:14splitting them out based on the amount

49:16of items that are in that array. So, let

49:19me just show you. Let me just run this

49:20right here.

49:21And let me show you that now we have

49:23five items.

49:24So, as you can see here, it went from

49:26one item, which is the array, to five

49:27items. And so, what we're doing here is

49:29we're processing each person's details

49:31individually through the automation in

49:33the next steps. And this is splitting

49:35out. So, this is when you have an array

49:37and you want to split out each item of

49:39that array. Now, aggregate right here is

49:41the exact opposite. As you can see, the

49:43diagrams are opposites. And so, what we

49:45do here is just bring it back to the way

49:47it was before.

49:48Right? If I press execute step, I can

49:50see that now I have the array that is

49:52back there, which is one item.

49:54So, you can see we have one item, five

49:56items, and one item as well. Now,

49:57typically, you would use split out

49:59whenever you have cuz a lot of the times

50:00we get data

50:02all within the same array.

50:04We get a bunch of details, contacts, and

50:05so on, all within one singular item. And

50:08so, what we want to do is go inside that

50:10item and take them out of each one and

50:12process them individually. In case you

50:14want to add them individually to, let's

50:16say, a native data table or even a

50:18Google Sheet. Right? You have to process

50:20each one individually for it to actually

50:21make sense to add them to a database or

50:24whatever it is that you want to do.

50:25Whilst this node right here, basically

50:26when we have a lot of information that

50:28is a lot of items, and we want to put

50:30them all together, smash them all

50:31together, to then send them through to

50:33the next steps. Right? So, it's

50:34aggregating data, so it's not

50:36everywhere. It's not in different

50:37places. It's all within one place. Then,

50:39we get on to the logic, which is if. So,

50:42this is saying, "If this equals to this,

50:44then it's true. If not, it's false." So,

50:47if I say, in this case, let me just uh

50:50take this out. Uh there we go, right

50:52here. So, let me say, "If the name

50:55contains um contains

50:58Sarah,

51:00which it only be one name, then we send

51:02it through the true route." Right? So,

51:04if I go here, I can now see that only

51:06one went through here, and the other

51:08ones went through here. Now, this is

51:09great because automation, again, is

51:11logic. So, if you want to send the

51:13automation in different ways based on

51:15whether something is true or false, then

51:17you can use the if node right here. Now,

51:19the switch node is a bit different

51:21because it gives you more flexibility.

51:22Right? Let me show you exactly what I

51:24mean.

51:25Let me say right here, inside, I have

51:28options to basically root the

51:30automation. So, I can say,

51:32"If the name

51:34equals to Sarah

51:36Johnson,

51:37if the name equals David Smith,

51:39then we can rename the output to Sarah,

51:41rename the output to David."

51:43And what this will do is that it will

51:45run through every single one.

51:47And the ones that are Sarah,

51:49or Sarah Smith in this case, Sarah

51:50Smith, Sarah Johnson, it will send them

51:52through Sarah.

51:54And if it's David

51:55Smith, then it will send it through

51:56David. So, if I run this, I can see that

51:59we only have one item and one item here

52:01because those are the only ones that

52:02applied to the logic. And so, how is

52:04this different from if? Well, if only

52:06gives you the opportunity to add

52:08conditions, but don't add more than true

52:11or false options. Whilst here, you can

52:13put as many options as you want. And

52:15this is great because you could have a

52:17use case where you have emails. You can

52:19categorize the emails based on whether

52:20they're FAQ, whether they are uh

52:23promotional, whether they are just

52:24normal, whether they are something else.

52:26And an AI categorizes it, and then you

52:28can use a switch node to then say, "If

52:30the email is FAQ, then you send it this

52:33way. If it's something else, you send it

52:35this way. If it's something else as

52:36well." Because you can keep adding root

52:38rules. You can say,

52:40"If the this equals to whatever, just

52:42hello."

52:44You can rename it to FAQ.

52:46And this will be another option, right?

52:47And you can do as many options as you

52:49want, which is amazing because that

52:50allows us to be able to root the

52:52automation based on [music] the type of

52:54input or the output that we get in the

52:56previous step. All right, then we have

52:57the code node. Now, the code node is

52:59amazing uh because it allows us to

53:02be able to do a lot of things at once.

53:04And code is one of the things that,

53:06again, a lot of no-code platforms don't

53:07have because in the name no code. Uh but

53:10the fact that n8n added the code option

53:11here made it so much easier for us to be

53:13able to to transform information from

53:16unstructured to structured in the

53:17easiest way possible. And it's very,

53:19very fast. Yeah, so let's say we have

53:20the array here, which is the exact same

53:22as we had before, which are just a bunch

53:24of names that are all uh within the same

53:27array, and we can't process each one

53:28individually now. What we can do with

53:30the code node using code, and by the

53:32way, if you're asking, "How did you

53:33write the code?" I just said, "Split out

53:35the items from the array." And the array

53:37is this. And then, if I run this, I can

53:39see that now,

53:41really quick, it uh split out the items

53:43in five different um separately. So,

53:45then we can go through five items

53:46individually. Now, this did the same

53:48function as this, so this wouldn't make

53:50a lot of sense. But, if you have

53:51something a lot more complex, and you

53:53need the data to be transformed in a way

53:55where it would take you seven or eight

53:58different nodes, then you can do it all

53:59within the code node. Now, I'm not going

54:01to get into detail about this because,

54:02again, we're all here to not write code,

54:04right? Uh but just know that you have

54:06this option to be able to write code

54:07with n8n, and it makes it a lot easier

54:10to do more complex stuff. You won't

54:11really need them for most of the time,

54:13but just knowing that you have the

54:14option there to use gives us a lot more

54:16flexibility when we're building

54:18automations. And then, we have the merge

54:19node. So, let me actually set this up.

54:21All right, so for the merge node, what I

54:23did is I put two different fields. One

54:25is hello, yo, and then the other one is

54:26hello, ciao.

54:28And when I run this automation, it will

54:30run this one first, and then it will run

54:32this one. And let's say I put a wait

54:34node here.

54:35I can put two 2 seconds just for it to

54:37wait 2 seconds. I can show you that

54:39this right here will run first, and then

54:42it will run this one. Right? Now, let's

54:44say that we had an automation where we

54:45had to run this, and then we wanted to

54:48run this, and then we wanted to go,

54:50right? Without us having to go through

54:51here first and run your whole

54:53automation, and then run through here

54:54first. Now, let's say that we had an

54:56automation that we start here, and it

54:58goes to different ways, but then we want

55:00to aggregate the data. We want to send

55:02the data up in the same way as two

55:04different inputs, input one and input

55:05two, to then be able to take both pieces

55:07of information to then be able to go to

55:10the next steps. So, that's exactly what

55:12the merge node does. If I go here,

55:14I can see

55:15also I can add more inputs. So, if I

55:17press four, it basically uh it can it

55:19can intake more inputs at the same time.

55:21In this case, I can do two. I can

55:23execute the step, and now we can see

55:26that we got two items. One is yo, and

55:28one is ciao. So, what is doing here?

55:30It's it's aggregating both of them

55:32together. Now, there are other options

55:33as well um to combine, to then SQL

55:37query, so you can query different things

55:39or different data points. And then, we

55:40can also choose a branch. Now, honestly,

55:42the one that we're mostly going to use

55:44is going to be combine, and it's going

55:45to be append. Now, append in this case

55:47would be output each item individually,

55:50and combine is taking these two and just

55:52putting them all in the same sort of

55:53array that you can use for the next

55:54steps. Now, here's a system that I built

55:56that uses the merge nodes, three

55:58different ones. Uh and the way that with

56:00this works is that we have um

56:02the LinkedIn post, which is made, the

56:04Facebook post is made, and then we want

56:06to merge them. So, we have the Facebook

56:08post and the LinkedIn post. And then, we

56:09have the Twitter post, which is made,

56:11the article, which is made, and then

56:12we're merging these two together as well

56:14to then finally merge all four things

56:16together so we can directly send to the

56:19Google Sheet. And the reason why we're

56:20merging them is because if we run each

56:22one individually, then we would have to

56:24set up a new automation for each row at

56:26a single time to add them to the Google

56:28Sheet, which we can actually just remove

56:30by just merging them all together in the

56:32same sort of structured way, different

56:34items, to then send them to the Google

56:35Sheet. So, the next nodes right here are

56:37going to be connectivity and API. Now,

56:39API is the most important skill when it

56:41comes to automation. And if you haven't

56:42watched my video about fundamentals of

56:44API, just watch it up here. Um but it is

56:46the reason why apps get to talk to other

56:48apps. Right? It's the reason why

56:50this app right here

56:52can talk to n8n.

56:53But this app right here can talk to this

56:55one, right? Or we can just talk to each

56:56other um in different ways. So, in this

56:58case, the first node that we have to

57:00learn is the HTTP request. Now, this is

57:02one of the most powerful nodes in n8n

57:05because when we go here,

57:07and I go to action app, so this is

57:09taking an action within a single app, we

57:11can see that we don't have an infinite

57:12amount of nodes. Right? Like we can stop

57:14here, and then what? What if you have a

57:16software that we want to use, but we

57:18can't find it here? What do we do then?

57:20Well, in this case, we use the HTTP

57:22request. So, let's say I wanted to use a

57:24free weather API, which gave me the API

57:26of the weather.

57:27And if I go here to n8n,

57:29I can't see any app that's called free

57:31weather API.

57:32So, what I have to do is I have to set

57:34up the HTTP request all to the free

57:37weather API, which I can find in the

57:39documentation right here. By the way, I

57:40covered this in the API fundamentals

57:42video. So, if this makes no sense to

57:43you, please go watch that video cuz it

57:45will help a lot. Um but this right here

57:47is the API documentation, which is

57:49essentially a documentation which allows

57:50you to see exactly what you can automate

57:52and what you cannot automate within a

57:53software, and also how to set up the

57:55automation. And so, by looking at this

57:57documentation here, I was able to set up

57:59the um the automation of the weather

58:01API, which looks at the temperature in

58:03London right now. And if I execute the

58:05step, I can see that now I'm still able

58:07to actually run the automation. I was

58:09still able to use the software without

58:11having to use

58:13the softwares here, which are already

58:15pre-made by n8n. And you can see here we

58:17have the different outputs. So, London,

58:19[music] we have the temperature degrees

58:20in Celsius, Fahrenheit, and different

58:22other informations as well. Then, we

58:24have webhooks right here. Now, webhooks

58:25are a way for us to get notified when

58:27something happens. So, let's say someone

58:29fills out a form,

58:30well, it sends the data to the webhook.

58:32You remember this right here, the

58:33Typeform trigger? Well, behind the

58:35Typeform trigger is actually a webhook,

58:37which allows us to be able to get

58:38notified when something happens. Well,

58:40in that case, it's whenever the form is

58:42being submitted, it sends the data to

58:44the webhook through an API. Right? So,

58:47right here we have the test URL, which

58:48is used for testing, and production URL,

58:50which is used when the automation is set

58:51to active. And then, we have the

58:53different type of HTTP method, which,

58:55again, is very similar to here because

58:56we have post, we have all these options,

58:58and these are the same ones that we have

59:00here. And then, we have the path, which

59:02is just the way that we that we name the

59:03URL. And then, this right here is the

59:05thing that you would connect to the

59:07software itself to then be able to for

59:08us to get notified when something

59:10happens within a software. And respond

59:11to webhook will be the thing that will

59:13then respond back to the server that we

59:15got the information from. Now, to show

59:17you exactly how it works, what I'm going

59:18to do is use a software called Postman

59:21API, which allows us to send example uh

59:24test data to the webhook, right? I'm

59:26going to go in here, copy this. Make

59:28sure you change this to post request,

59:29and then I'm going to paste this here.

59:31Make sure this is a post request as

59:33well. And then the body will be what is

59:34the thing that we're sending the

59:35webhook. In this case, let's just do

59:37full name.

59:39My name.

59:40Let me run this here.

59:42I'm going to press send.

59:44This will now start the workflow, as you

59:45can see here. Now we started the

59:47workflow and we sent the full name here,

59:49which is the body, right? And this is

59:51how we send information from one server

59:53to another through a webhook. [music]

59:55Now let's say I had this connected. Now

59:57all I have to do here is I have to

59:58change this to using respond webhook

1:00:00node.

1:00:02And here,

1:00:03um, what we can do is

1:00:06we can just respond with

1:00:08workflow has finished.

1:00:13And so when I run this

1:00:15I send this here.

1:00:17I can see that now we get the data,

1:00:18workflow has finished.

1:00:20Now this is amazing when we are sending

1:00:22data to a software

1:00:24or to a webhook in this case and we want

1:00:25to get something back.

1:00:27So let's say we send the sign-up link to

1:00:29a user in some sort of software. He

1:00:31signs up, we get something back. Lastly,

1:00:34it would be the AI integration. So this

1:00:35is one of the ones that are most

1:00:37commonly used. First one is the AI node.

1:00:40So this is just to create something with

1:00:42AI. So in this case, we can say create a

1:00:45LinkedIn post

1:00:47about

1:00:48life. And here we have our OpenAI

1:00:50connected. The resource is text. The

1:00:52operation, which is the action that

1:00:53we're taking, is good. The model, it can

1:00:55be You have a list of models here.

1:00:57You can use GPT-4 or latest, or you can

1:00:59use whatever you want. And then you have

1:01:01the prompts right here. The prompts,

1:01:02there are three different prompts that

1:01:03you can write. The first one is the

1:01:05system prompt. The system prompt is a

1:01:06prompt that you tell the AI you give it

1:01:09an identity. So here you are a helpful,

1:01:10intelligent XYZ assistant. Then you have

1:01:13user. So user is when you tell it to do

1:01:16something. So your task is to do XYZ. An

1:01:18assistant [music] is when you give it

1:01:20some examples. In this case, let's keep

1:01:21it simple. I'm going to press execute

1:01:23step. And this will now create a

1:01:25LinkedIn post about life. As you can

1:01:27see, sure, here's a thoughtful and

1:01:29professional LinkedIn post about our

1:01:30life. And then it gives me a whole

1:01:31LinkedIn post that we can then use. Now

1:01:33this AI step is used whenever you want

1:01:35to have some sort of data or some sort

1:01:37of input sent to AI, just like we would

1:01:40on ChatGPT, automatically and then give

1:01:42us the output. For content, for any

1:01:44structured data that you want

1:01:45structured, for anything pretty much

1:01:47that you can think of AI playing a role

1:01:49in the automation, just as a linear

1:01:51thing, for it to do just that task in

1:01:52specific, only that task. And then we

1:01:54have the AI agent. If you haven't

1:01:56watched my AI agent 101 video, make sure

1:01:58to watch it up here. It explained

1:02:00exactly how you can build your first AI

1:02:01agent from scratch using n8n. But the AI

1:02:03agent is the reason, probably the reason

1:02:05why n8n blew up so fast is because it

1:02:08had now the opportunity for us to use

1:02:11something that was connected to AI that

1:02:13could think through or that could

1:02:14remember the different conversation we

1:02:16had but that was also hooked up to

1:02:19different tools. So in this case, if you

1:02:20scroll down we can see that we hook it

1:02:22up to Gmail. We can hook it up to

1:02:24Airtable. We can hook it up to or pretty

1:02:27much anything.

1:02:28We have all these different softwares

1:02:31that we can use to then be able to

1:02:32connect it to the AI agent for it to

1:02:34actually take action. And the good thing

1:02:36is that the Gmail tool

1:02:38can be connected and the Airtable tool

1:02:40can be connected and the let's say

1:02:43Notion tool can be connected

1:02:45and you can have as many connections as

1:02:46you want and this acts as a personal

1:02:48assistant that allow us to be able to

1:02:51take action on different things based on

1:02:53the input that we give it. So in this

1:02:55case, what it would look like is we

1:02:57would have a trigger, which would be

1:03:00usually it would be an on chat message,

1:03:02which means that we're chatting with the

1:03:03AI agent itself. If I open chat, I can

1:03:05now see that I can speak to the actual

1:03:06AI agent, something that you can't do

1:03:08with the AI step. So I can say, "Hello."

1:03:11What this will do is that it will then

1:03:12talk to its AI. It will remember the

1:03:14conversation and then bring it back.

1:03:16It's sort of like a person that we're

1:03:17talking to. And the good thing is that

1:03:19the person itself

1:03:20has access to our softwares.

1:03:23Which is amazing because now we can take

1:03:24action on our behalf based on the input

1:03:27that we give it. So if we say draft an

1:03:28email, it will then go to the Gmail tool

1:03:30and then do its thing. Same thing with

1:03:32Airtable and same thing with Notion. And

1:03:33the use case that you can think of this

1:03:35is pretty much anything, any task that

1:03:37you want a singular input data store. So

1:03:40you basically have, let's say, one chat

1:03:42message, right? We have one place and

1:03:44through a series of inputs, so through a

1:03:46series of sending Gmail, sending

1:03:48Airtable or create a Notion task or

1:03:50whatever it is it takes actions in

1:03:51different softwares without us having to

1:03:54create a new automation for every single

1:03:56action that we want to take.

1:04:00In this video, I'm going to walk you

1:04:01through everything that you need to know

1:04:03about data transformation and using JSON

1:04:06within your automations. Now I know this

1:04:08topic scares quite a few people and it

1:04:10scared me coming into automation as a

1:04:12beginner with no technical knowledge,

1:04:13but it actually is quite easy and very,

1:04:16very important to understand. All right,

1:04:17so JSON 101. JSON stands for JavaScript

1:04:21Object Notation. So if you see

1:04:22JavaScript here and there, it stands for

1:04:24JSON cuz it's JavaScript, so J S O N

1:04:27right here.

1:04:28What it is, it's a way for servers to

1:04:32store and transfer data. Right, it looks

1:04:34like text, but it's structured so

1:04:36computers can read it. Now the best way

1:04:37that I can describe JSON is English for

1:04:40computers. It's the language that

1:04:41computers use to speak to each other.

1:04:43Right? And that's why it's very, very

1:04:45important for us to understand it

1:04:46because pretty much everything is built

1:04:49in JSON. Right? APIs, which I spoke

1:04:51about in another video up here. Config

1:04:53files, databases, but also your favorite

1:04:55platforms like make.com, n8n or Zapier,

1:04:58they're all built on JSON. It's crazy to

1:05:01me that people still don't understand

1:05:02that cuz behind the nodes, the squares

1:05:04and the [music] the steps if you really

1:05:06look at it, it's all JSON. That's why

1:05:08it's very, very important to understand.

1:05:09Now we use a very basic structure in

1:05:12JavaScript. [music] When we write JSON,

1:05:13we use key-value pairs. So the keys are

1:05:16strings. They're always going to be in

1:05:17quotes. And the values can be string,

1:05:19can be my name, number, can be an age,

1:05:22boolean, which is true or false, null,

1:05:24when you get no data, it's going to be

1:05:25zero, zero data, object, which is

1:05:27basically a JSON inside the JSON, an

1:05:30array, which is a list of items, which

1:05:32I'll walk you through step by step. So

1:05:33these are all the different types of

1:05:35values that you can get using the key.

1:05:38Here's an example of a basic key-value

1:05:40pair. We have the key, which is name. We

1:05:42have the value, which is Michele in

1:05:44quotes because it's a text string.

1:05:46The age, which is key, which again is in

1:05:48quotes because it's always a string. The

1:05:51age, which is not in quotation marks

1:05:52because it's an age, so it's a number.

1:05:54Then student and then true. Which is

1:05:56boolean, which is true or false, right?

1:05:58So we have again name, key, key and key,

1:06:01and these are values, but they're

1:06:01different types of values. Now arrays

1:06:03are ordered in list inside of the

1:06:07brackets. Now let's say you go to the

1:06:08grocery store and you get the groceries.

1:06:11Right? The groceries isn't one thing.

1:06:12Groceries is a list of items inside the

1:06:15key, which is groceries. And so this is

1:06:17the same exact example when we have

1:06:19skills. So the key is skills and the

1:06:22value is automation, AI, Notion, which

1:06:24are my skills. And so why do we use

1:06:26these? Well, we use this when we have

1:06:27the same key, like groceries or skills,

1:06:29but we want to add multiple items in the

1:06:32array, right? The array is just the the

1:06:34list of things,

1:06:35um, to to explain the key, right? To

1:06:37explain skills, to explain groceries, to

1:06:39explain whatever it is.

1:06:40We have nested objects, which are JSON

1:06:43inside the JSON. And for example, we

1:06:44have the key, which is person. The key

1:06:47is also name, but the name is inside the

1:06:48person, so it's JSON in JSON. Um, and

1:06:51then we have contacts, which is email

1:06:53and phone number. So we have different

1:06:55pieces of information, but it follows a

1:06:57very clear hierarchy. Hierarchy just

1:06:58means what is the first thing that you

1:07:00should see and then what is the next

1:07:01thing and what is the next thing. Of

1:07:02course, the first thing is person. The

1:07:04next thing [music] is name. And we have

1:07:05contacts within the name and we have

1:07:07email and phone number. And then we

1:07:08close off with three curly brackets

1:07:10because we start with one, two and

1:07:11three.

1:07:12Right? So nested objects is usually the

1:07:14thing that we always [music] use and

1:07:16that n8n, make.com usually uses when

1:07:19they want to show different types of

1:07:20data inside the data. Right? So you

1:07:22might have a Google Sheet and the Google

1:07:24Sheet might show the ID of the sheet. It

1:07:26might show the sheet name. It might show

1:07:28the row values. It might show different

1:07:30types of things. And so to show those

1:07:32things, it uses the

1:07:34nested objects. And so JSON applying

1:07:37into APIs, and if you've never heard

1:07:39APIs, it's just basically how a software

1:07:41speaks to a software, how these

1:07:42softwares are actually able to work

1:07:44through APIs.

1:07:45This is a HTTP request, which is a

1:07:48request that we send to a server to do

1:07:50something. So in this case, we have a

1:07:51curl and a curl is just a bunch of JSON.

1:07:53Now don't worry about what this means,

1:07:55but this is a request. This is something

1:07:57that we send to the server saying, "Hey,

1:07:58can we get something? Can we Can we do

1:08:00something?" Right? In case you want to

1:08:01get new rows or you want to add a row to

1:08:03a Google Sheet or you want to add a

1:08:04contact on ClickUp or do something else,

1:08:07this is what you use. And then the

1:08:09response, which is always going to be in

1:08:10JSON, looks like this. So let's say

1:08:13we're looking for users in a software,

1:08:15let's say. And the output that I get is

1:08:17the ID of the user, the name of the

1:08:18user, the email, whether the user is

1:08:20active or not and [music] the roles of

1:08:22the user, which is an array because

1:08:24there can be different roles within the

1:08:26same person. And so we have the keys,

1:08:27which are all inside the quotation marks

1:08:29because they're all strings.

1:08:30>> [music]

1:08:30>> And we have the values. But the values

1:08:32can change based on whether it's a

1:08:33number, a string like this

1:08:36a boolean or just an array. So how does

1:08:39this now apply to n8n? So if I go to

1:08:41edit first step, I can add a manual

1:08:43trigger, which is just a first step of

1:08:45the automation manually. And let me add

1:08:47another node, which is edit fields.

1:08:50Now this node right here is very, very

1:08:52important and is something that we

1:08:53typically use to transform data, which

1:08:56is getting an input, which can be a

1:08:57name, it can be a bunch of values, then

1:08:59putting it here right inside the edit

1:09:02fields node, which transforms the input

1:09:04and then gives us the output. So if I go

1:09:05here to mode, I can see that I can do

1:09:06manual mapping. Let me zoom in. Or I can

1:09:09actually write things in JSON. And this,

1:09:12again, is the key value pair, key value,

1:09:15which is the exact same as the thing we

1:09:17spoke about. Now if I go to manual

1:09:18mapping, I can now start adding fields

1:09:20to set. So I can say my full name

1:09:23is Michele

1:09:24Sorti.

1:09:25Right? And I can see that this is a

1:09:26string. Why is it a string? It's because

1:09:28it's text. If I put number and I run

1:09:31this,

1:09:32this would be null because there is no

1:09:33number. It's just text. And it deletes

1:09:35it, right? And so, if you go back to

1:09:37string, put my name,

1:09:39what I can do here is execute the step,

1:09:42which will show me the output in schema,

1:09:43which is basically the way that we all

1:09:45look at it because it looks a bit messy

1:09:47in JSON, the table version,

1:09:49or JSON, right? So, if you've never

1:09:51pressed on this button JSON, this is

1:09:53what it looks like. We have an object

1:09:56right here. We have the key and we have

1:09:58the value

1:09:59right here.

1:10:01Now, let's say I wanted to do age

1:10:03and we do um 20.

1:10:07If I execute the step, now you can see

1:10:09that the object right here is increasing

1:10:12in the size because there's a new

1:10:14key-value pair. And bear in mind that

1:10:16this is text right now. It's not a

1:10:17number. So, if I wanted to change this

1:10:19to a number,

1:10:20look at the color of this, it changes

1:10:22the actual type of output. Right? So, it

1:10:24changes the type of output that comes

1:10:26here to 20. Let's say I wanted to do

1:10:28Boolean.

1:10:29This is true or false.

1:10:31I can say student

1:10:34false or you can do true.

1:10:36So, we're saying, "Hey, [music] his name

1:10:37is Michele Torti. He's 20 years old and

1:10:39he's a student, right?" And then you can

1:10:41also be adding an array, which in this

1:10:43case can be skills, like I mentioned

1:10:44before, and you do

1:10:47square bracket

1:10:48square bracket to end it. Then you put

1:10:50skills. You say notion.

1:10:53You finish it off, then and then done.

1:10:56Then make

1:10:58and then you do something else, right?

1:10:59But we are adding items inside the

1:11:01array, which is here and here. Two items

1:11:05right here. Boom.

1:11:07Now, if I press execute step, this will

1:11:08now show the skills like this. It will

1:11:11show the full name, it will show the

1:11:12age, the student, and the skills right

1:11:14here, which are wrapped up in square

1:11:16brackets right here.

1:11:18Notion, make, and so on.

1:11:19And finally, we have

1:11:22the object, right? And the object can be

1:11:25a summary. Let's call it summary and

1:11:26let's copy this full thing, this whole

1:11:28object, because this is an object again,

1:11:31from here to here.

1:11:33And I can paste this here.

1:11:35If I execute the step,

1:11:37this will now

1:11:38add the new object that we have, which

1:11:40is called summary,

1:11:41and we have it this again,

1:11:43which we just copied up here. Now, why

1:11:45am I showing you this? It's because

1:11:46typically when we get inputs from the

1:11:48previous node, the previous step, we

1:11:50send it here to then be able to change

1:11:52the data. So, let me show you an example

1:11:54here. Let's say I have Let me delete

1:11:55this. Let's say I have the full name and

1:11:57I wanted to get just the first name.

1:12:00So, let's say I run this

1:12:02and I have the output, which is full

1:12:04name. Let's say I add another edit node

1:12:07right here. I can go here. I can remove

1:12:09[music] this. And now, let's say I

1:12:11wanted to take the full name and I

1:12:12wanted to transform it into just the

1:12:14first name. What I do here is I press

1:12:16here. I can say first name.

1:12:19And if I go here to expression,

1:12:22I can put curly bracket curly bracket.

1:12:25And now inside here, it says anything

1:12:27inside the curly bracket curly bracket

1:12:29is JavaScript, which is JSON. Right? So,

1:12:31we're saying I can actually reference

1:12:33previous variables using this. So, I can

1:12:35say dollar sign JSON, which is the way

1:12:38that you start when you want to

1:12:39reference some sort of item here,

1:12:41dot

1:12:43full name,

1:12:44right? Which you got right here, full

1:12:46name. And so, that is how you can

1:12:48reference these variables from the

1:12:49previous steps to the next steps. And

1:12:52that's how dynamic variables work,

1:12:53right? Because they are dynamic, they

1:12:55change every single time. Cuz right

1:12:56here, if I run this, I would also have

1:12:59the first name, which is Michele Torti.

1:13:01But if I go back here

1:13:02and change it to James,

1:13:06and I run this again,

1:13:08this will now be James, right? Which is

1:13:09why it's dynamic, which is why

1:13:11automations are brilliant because they

1:13:12change data and you can automate the

1:13:14process so you don't have to do it

1:13:16yourself.

1:13:18So, okay. Let's say we have the full

1:13:19name here. Let me go full screen by

1:13:21pressing this button right here.

1:13:23And let's say I wanted to take this

1:13:24name. I just wanted to get James. Now,

1:13:26inside N 10 and the reason why I would

1:13:29say for a beginner it looks a bit

1:13:30technical uh because of this, because

1:13:32make.com doesn't actually show you this,

1:13:35but that's what it is in the back end.

1:13:37I can, after the whole variable, so full

1:13:39name, which again is this,

1:13:42what I can do is actually use some

1:13:43formulas. So, I can put dot, which now

1:13:46allows me to see different things that I

1:13:48can do with this full name. I can split

1:13:51the first name, which we're going to do.

1:13:52I can check whether the um the name

1:13:54includes something, right? So, you can

1:13:56see here whether the team includes T and

1:13:59it says true. And so, these are all

1:14:00formulas that you can use within your um

1:14:04the full name. Honestly, you'll probably

1:14:06use maybe five or six of these, so don't

1:14:08even worry about understanding each one.

1:14:10Um one of them that I use a lot is split

1:14:12because I usually want to split the full

1:14:14name so I can get the first name. So, in

1:14:16this case, what we're doing here is

1:14:18we're getting the full name, which is

1:14:19James Torti. What we're doing is saying,

1:14:21"Hey, let's split this name by the

1:14:23presence of a space so it gives me this,

1:14:26which is first name, and then last name.

1:14:29And then we want to get the first,

1:14:31first one, which is James.

1:14:34And that's the way that this works

1:14:35because everything in JSON is logic.

1:14:37Everything in code is logic. Right? So,

1:14:39if you put logic here into this

1:14:41automation, or not automation, in this

1:14:43variable, then you're saying, "Hey,

1:14:45let's split

1:14:47this by the presence of a space." So, we

1:14:49do the apostrophe here and we actually

1:14:52put the space in our keyboard. So, now

1:14:54we can see that we have the first one

1:14:56and we have the second one. And now,

1:14:58after we put split, then we want to put

1:15:00dot

1:15:01first

1:15:03because we're returning the first

1:15:04element of the array. Cuz you can see

1:15:06here, this is an array,

1:15:07right? Because it has this {comma} this

1:15:10{comma} that, right? So, let's say I had

1:15:12the full name being James Torti Arthur,

1:15:15it would then be James {comma} Torti

1:15:16{comma} Arthur. And so, after this, we

1:15:18want to put first

1:15:19>> [music]

1:15:19>> and this is how we get James. And that's

1:15:21how in emails you can turn the full name

1:15:23into a first name by using JSON and by

1:15:26using formulas within the JSON. Now, you

1:15:28looking at this right now, it probably

1:15:29looks very overwhelming and it was very

1:15:31overwhelming for me at the start as

1:15:33well. But when you understand the ones

1:15:35that actually do matter, cuz yeah, let's

1:15:37be honest, you're not going to be using

1:15:38all these,

1:15:40they work very well, right? Like if you

1:15:41understand split, for example, and

1:15:43understand that this is the formula that

1:15:45you use every single time to split a

1:15:47full name from something else, then

1:15:49you're pretty much set. That's it,

1:15:50right? And the good thing is that you

1:15:52have an AI

1:15:53on the right-hand side

1:15:55that you can ask any questions to, which

1:15:56can also help you. And so, that is how

1:15:59we can transform data. We can take some

1:16:01sort of full name, we can then use some

1:16:03formula, and then we can output the

1:16:06first name, right? From here

1:16:08to here.

1:16:09And that is data transformation. That's

1:16:11how we use JSON to be able to take an

1:16:13input and turn it into something that's

1:16:15an output. And quick side note here, if

1:16:17you're wondering what's behind this, I

1:16:18mentioned it's JSON, let me show you

1:16:20what I mean. I can copy this. I can go

1:16:22here and if I paste this, this will be

1:16:24JSON because these are nodes that have

1:16:27parameters, that have different stuff,

1:16:29which are all JSON. And the reason why

1:16:32this N 10 make.com works very well is

1:16:34because we don't have to see this. But

1:16:36just know that behind this

1:16:38is this, right? So, it's all JSON.

1:16:40That's why understanding JSON is very,

1:16:41very important because it gives you the

1:16:42understanding of how things actually

1:16:44work together and how the English

1:16:46language in computers actually works.

1:16:47Now that we have this, full name

1:16:49>> [music]

1:16:49>> and then going to the first name, now we

1:16:50can add some branching. So, we can add

1:16:52some some conditions.

1:16:54And we can add if. So, if is saying,

1:16:57"Hey, if this variable,

1:16:59which again we can get by doing curly

1:17:01bracket curly bracket,

1:17:03JSON dot first name.

1:17:06So, JSON dot the variable that's here,

1:17:08right? JSON dot first name

1:17:11equals to James, this will say true,

1:17:14true branch. And if I go outside of

1:17:15here, I can see that this is a true

1:17:17branch. This is a thing that is true.

1:17:19Then we can now send the automation one

1:17:21way if that specific criteria or

1:17:23condition is true. And if it's false,

1:17:25then we go down this way. And for

1:17:26example, if I change the the name to

1:17:29Michele,

1:17:30execute the step, now this will be a

1:17:31false branch. This will go down

1:17:33defaults. And again, behind this, what

1:17:35it does is that it uses this variable.

1:17:38It then says, "Hey, is this variable

1:17:40equal to this? If yes, [music] then it's

1:17:42true. If not, then it's false." Because

1:17:44again, everything lies on logic.

1:17:46Everything is logic in automations.

1:17:48Whether you have this or whether you

1:17:50have [music] this, conditioning into

1:17:51different paths. All right. So, right

1:17:52here I have an AI agent. So, we'll see

1:17:54how JSON actually applies within the AI

1:17:56agent. I can open chat and I can say

1:17:59hello.

1:18:01And what it's doing now is it's

1:18:04is it sending the information here,

1:18:06JSON, which is session ID, action, and

1:18:09chat input, which is the thing that I

1:18:10just told it to do. So, these are

1:18:11different items that we get as an output

1:18:13from the actual thing. It's transferring

1:18:15data, so what you see here, the line is

1:18:17actually just transferring the data

1:18:18here,

1:18:19to this AI agent, which is giving the

1:18:21chat input as an example in JSON here.

1:18:24Then,

1:18:25using the chat input, is then talking to

1:18:28the actual open AI by giving it JSON

1:18:30again, which is messages,

1:18:32right here.

1:18:33And then the output that we get using

1:18:35open AI is this. "Hello, how can I

1:18:38assist you today?" Which again is the

1:18:40output, which goes here. "Hello, how can

1:18:42I assist you today?" So, everything that

1:18:43you see within your automations, data

1:18:45here and data there, or some sort of

1:18:47variable going to the AI agent and then

1:18:48going to something else, it's all JSON.

1:18:51That's how data flows within different

1:18:52automations. And so, let's say now I

1:18:54wanted to take action on my email. I can

1:18:56say send an email

1:18:58to

1:18:59michele.torti5@gmail.com

1:19:01saying we have

1:19:04uh

1:19:05football tomorrow. I can press go. This

1:19:07will now send the data here. It will

1:19:08then call the tool

1:19:10and then it will send the email, right?

1:19:12And if I go inside here, I can see that

1:19:14this is the input. The messages is the

1:19:16input, the subject line and email body.

1:19:19And the output is the response, which is

1:19:22the email ID that I just made, the

1:19:24thread ID which is a thread of the

1:19:25email, and then the label ID telling us

1:19:28that this is sent. And this right here

1:19:30is an object. Right right here. The same

1:19:32way that it was person name, it was

1:19:34contact, it was whatever it is. All

1:19:36right, and lastly here we have a form

1:19:37that I just made.

1:19:39If I go to the form, I can see that I

1:19:40have the full name, the age, and the

1:19:41date. So I say Michele Torti.

1:19:44The age will be let's say 20.

1:19:46The date, let's say the date is today.

1:19:48The 20th. I can press submit.

1:19:51What this will now do is it will send

1:19:52the data to our automation.

1:19:55And we get back the full name which is a

1:19:57key

1:19:58value.

1:19:59Key, as you can see here this is a value

1:20:01but it's not in quotation marks because

1:20:02it's a number. And then everything else

1:20:04is a key value pair with this because

1:20:06it's all a string. It's all text. Now

1:20:08some fundamental rules to remember for

1:20:10JSON is that the keys must always be in

1:20:12double quotes. The keys right here,

1:20:14name, must always be in these two codes

1:20:16because it's always a string. If you

1:20:19don't put it that way, it'll break,

1:20:21right? The second thing is to understand

1:20:22the difference between string, number,

1:20:24boolean, null, object, and array, right?

1:20:27And how these sort of function because

1:20:28then you're able to understand JSON as a

1:20:30whole because it is just that. It is

1:20:32just strings which are text, numbers

1:20:34[music] which are just numbers with no

1:20:35quotation marks, boolean true or false,

1:20:37null no data, object which is JSON

1:20:40inside the JSON, and array which is a

1:20:42list of values. Then we have the third

1:20:44rule which is separating pairs with

1:20:45commas. So we have the key which is

1:20:47name,

1:20:48the actual value, comma, the next thing,

1:20:50comma, the next thing, and then the last

1:20:52one doesn't have a comma, the one

1:20:53nearest to the actual curly bracket

1:20:55which finishes off the JSON structure.

1:20:57And lastly is wrapping the whole thing

1:21:00in quotation mark if it's an object. And

1:21:02an object in this case is just a list of

1:21:04key value pairs, right? Key, value, key,

1:21:07value, key, value. And we wrap up the

1:21:09whole object using curly brackets here

1:21:12and curly brackets here to finish it

1:21:13off. And if you're someone who's serious

1:21:15about starting and scaling your AI

1:21:17agency, then you might want to check out

1:21:19the first link down below which is a

1:21:20video that walks you through our

1:21:21one-to-one mentorship program which is a

1:21:23program where we get to work with you

1:21:24one-to-one to help you start and scale

1:21:26your agency to 10K a month.

1:21:31Webhooks are by far one of the most

1:21:32powerful thing within automation. Yet

1:21:34most people still have no idea how they

1:21:36work. So today I'm walking you through

1:21:37exactly what a webhook is, how you can

1:21:39use it in real life, and how you can set

1:21:41it up within your own N and so by the

1:21:43end of the video you'll be able to use

1:21:45it within your own automations even if

1:21:47you've never touched it before. So let's

1:21:48dive in. So what a webhook is is

1:21:50essentially a URL that allows us to get

1:21:51notified when something changes within

1:21:53an app. So think of it as someone coming

1:21:55into your house and ringing your

1:21:56doorbell. As soon as they ring your

1:21:58doorbell, you instantly get notified,

1:22:00right? And we could usually refer to

1:22:02this as an instant trigger. That's

1:22:03exactly what a webhook is. As soon as

1:22:05something happens, we instantly get

1:22:07notified in the webhook. Now the

1:22:09business use case to this could be

1:22:11someone filling out a form. So as soon

1:22:13as the user fills out a form, whether

1:22:14it's in the website or something else,

1:22:16we instantly get notified so the webhook

1:22:18receives the data meaning receives the

1:22:19responses from the form and then starts

1:22:21the automation which is the one right

1:22:23here. All right, so let's jump onto N

1:22:24and right here real quick and we can add

1:22:26the first step. So adding the first step

1:22:27just asking us what is the trigger? What

1:22:29is a thing that starts the automation?

1:22:31As I mentioned, when someone fills out a

1:22:32form, the first step will be the webhook

1:22:34because that's the thing that allows us

1:22:35to actually start the automation. Now

1:22:37the webhook can be found here. So on

1:22:38webhook call, we can obviously trigger

1:22:39manually an app event on a schedule,

1:22:41form submission. We can do a bunch of

1:22:42stuff but the webhook itself is this one

1:22:44and it runs the flow on receiving an

1:22:46HTTP request. Now HTTP stands for

1:22:48hypertext transfer protocol, very fancy

1:22:51way of saying that data is passed from

1:22:54one app to another when we get notified,

1:22:55right? So when we tap on here, we get

1:22:58introduced to this page right here.

1:22:59On the left hand side is basically

1:23:01pulling in from the webhook so it's

1:23:02listening for new events. So when we

1:23:03press this button, it now the webhook

1:23:05itself it's listening for notifications

1:23:08coming from some server, some sort of

1:23:09server.

1:23:10So let's stop listening.

1:23:11And in the middle, this is the

1:23:13configuration and on the right hand side

1:23:15is the output. So when we get notified,

1:23:17the notification will show up here.

1:23:19So right here the first thing that we

1:23:20see is the difference between a test URL

1:23:22and a production URL. Now this is quite

1:23:24misunderstood between what they are but

1:23:26a test URL so if you look at the the

1:23:28actual thing, the actual URL in this