Transcripts
1. Class Introduction: Hi, and welcome to this
very quick tutorial for automating
your Twitter feed. We're going to complete this
whole process in 20 minutes, and I'll put a timer on that
we're going to stick to. The feed that we generate
comes from Google and should work on most topics
and is fully up to date. You'll be able to
set those updates and be able to post actually
as often as you like, daily or even hourly
if you wanted to. And I'll show you the
results in real time. Here's a feeder I've
created for my own channel, which is an AI music channel. All the entries, you can see, have a custom description,
relevant hash tags, and usually with an image too, the tweets then click through
to the relevant content. If you've got 20
minutes to spare, then I'm going to show
you how to do this. So I will see you in
the first lesson.
2. Grab Your Content Feed: Great. Thanks very
much for joining me. We're going to jump straight in, as we're going to try and
do this in real time. So I'm going to set
that 15 minutes, and we're going to try and
stick to that. Okay, let's go. First thing we do is we head to Google and google.com
slash ALERT. Okay, it's very, very
straightforward. In the creating alert about you put in the
topic that you want. It can be absolutely, you know, anything relevant to your niche or even something
new where you want create a news feed or something that complements
your existing niche. And we're going to
be able to run this once we've completed this
as often as you like, depending on the
content appears. So let me just have
a look at one. I'm clicking through
that headline. So this is the link
that will happen on Twitter when your users
click on any of those tweets. So that's one, let me come back. Let's try something else. I'm going to put in space race. And you can see that there's lots of news items
that are being generated by Google from all around the
world and it just clicks through to whichever news
outlet is providing that story. This is the photograph that will appear
actually on Twitter. Okay, my niche area is AI music, so we're going to use
this as my example. And what you do is you
go create an alert, where you'll see in
here, of course, I've already got one
done previously, but this is just to show you. You create an alert and then come over to
the pencil icon, click the pencil icon, and in the deliver to, you put RSS feed, and then you say update alert. Once you've done that, it brings you back
to the main screen and this icon here appears, which is your RSS feed. Now if you click on that, you can see this file, which you don't need
to worry about at all. It's an XML file. You take this URL, so you can copy this now
and put it somewhere safe or we'll come
back to this and we'll grab it again
for use a bit later. That is the first part, we're just a couple
of minutes in. Great. Let's move on.
3. Set-up ChatGPT API: The next part is to create
an account with CHAT GPT, but we need to be
able to use the API. So we need to do this
via the developer area. It's not difficult
to do. You don't need any technical background. Just go to platform dotpeni.com. And you sign in. Once that's done, you will come to a dashboard area
or your profile here. At your leisure, of course, you can have a good
look around here, but because we're on the clock, we're going to move forward. The only things that we
need to look at here are, first of all, in billing, you will need to add some money to your account in order to be able
to use this service. Let me say right here that depending because of the model that we're going to be using, you really will only use a few pennies
here or a few cents. It's not going to cost
very much at all. In fact, my balance there, I put $10 in and that was
I think over a year ago. But you'll see that shortly. Put in your credit cards
and give yourself maybe a $10 balance and
that now means that you can go ahead and
create your API key. If you can see my mouse, just click on API keys. I've already created one, but you simply will go to
create a new secret key. You can call it
anything. Twitter no. I'm just going to call
it my key for now. Create Secret key. You have a default project, you don't need to
worry about that. Create a secret key,
which it will do. There it is and you
will copy that key. Bear in mind here, it says, you won't be able to view this
key again. It's a one off. So once you've copied
that, then put it in a safe place
because that's what we'll be using a
little bit later. So copy your key, and I'm just going to
show you one last thing, which is my usage on here. So this shows you, you can come back to
this when you start using your Auto Twitter feed, and you can see
what you're using. We look here, I've only set this whole process up
just a week or so ago, and you can see, look,
it's using $0.01 there. There's an input, which is for Chat EPT of the prompt and
then the output response. So actually, even for both
of them, it's one set. So actually, all of January and with
the end of January now I've only used eight sets, so it really is pennies to
be able to use this process. There really isn't anything
else I need to mention here. Of course, this is a
detailed in depth area, and there's lots going on here. But for now, that's
all you need. Just make sure that you've
got your key, and again, we're just a couple
more minutes in, so we're going to move on to the next
part of the process.
4. Connect your X/Twitter Account in Buffer: Okay. The next thing then is to create an account
in buffer.com. Buffer.com is a social
media management platform, which is where we're going to be posting directly
onto the X platform, the tweets that we generate. Okay, if you haven't done
so, create an account. I'm going to log in. If you've not used it before, all we need to do is
one thing which is to connect our X channel to buffer. I can just show you I have
three channels over here. One is my channel. The other is an Instagram, and the other is a
Tik Top channel. So you get three for free, and you can actually even schedule a number of I'm not
sure exactly how many posts, but you can schedule a
number of posts to go out. But we're going to be using this service to post directly. So we come from make.com, and we just use buffer as an intermediary to
post straight out. Let's get this done.
So over on the right, click on channels because
you are going to need to connect your Twitter
X account to it. Just make sure you've got X open and logged
in in your browser, then you go Connect Channel. From Connect Channel,
you can see a list of the various social media channels that you
can connect to. You'll click here, Twitter
Stroke X. I should happen automatically and
you'll just approve your X feed so that
buffer has access to it. When you've done
that, that's it. That's actually
all we need to do. We don't need to linger around here because we
are on the clock. I'm going to just show
you a few seconds. That's all. You can skip this if you
don't want to see it. But just to show you that in
the calendar view over here, it should show you, yes. So these are my tweets that Buffer has been sending
out to the platform. If I just kind of
mouse over that one, you can see the content. It's, you know, the
content is literally a text description with some hash tags that
have been generated, and Twitter automatically finds. There's a buff URL as well, so Twitter finds an
image from that. Okay. That's buffer. Lots
to see in buffer, but not for us because
we're now moving on.
5. Putting it Together in Make: And now we are going to
join it all up in make.com. Right. This is the exciting bit. So again, if you're not
familiar with make.com, for now, create your account. I'm going to sign in. It's an incredibly
powerful tool, automation tool that connects so many different types of apps. It might be a little bit
overwhelming to start with, but you know, just take a look around and to really
be able to program it, you'll need a little
bit of experience, but for what we're going to do, I'm going to walk you
through it step by step, and it should work without
any problems for you. And then, of course, you
can take it further if you want to go into more detail. Okay. This is my dashboard, and it just shows the amount
of activity that I've, you know, had using make.com with the amount of operations
and data transfer. So an operation
is a single call. We are going to if you can
look at my active scenario, which is actually the one that
we're going to replicate, is this is the RSS feed. When I call the RSS feed,
that's one operation. That is my hatch EPT operation and that
is my buffer operation. We build up what
they call scenarios. For every operation, I
think it's one unit, and you can see you get
1,000 operations per month. Um, which allows you to do an
awful lot on the free tier. I think it's pretty generous. Right. Let's move on
and make this happen. What we're going to do now
is create our scenario. Let's do it. Create a new
scenario. Here we go. Right. You start off with a blank Canvas and big plus
button, which we click. Actually, we don't
need to click that. All we need to do is
search in the box here and we're going
to search for RSS. Watch the RSS feed item. First, you know, a
little module here, which is acts as a trigger for
the scenario to be called. And there's one parameter
that we put in here, and that's our URL. So do you recall the
URL that you saved? I'm just going to
grab another one. I'm going to grab it
on a different screen. The URL that you saved for your feed when you
created your alert, go and grab that and
put it in there. How many items do
we want to return? Ultimately, is we want one. We're going to create one tweet, based on the feed that we get. But because we're calling
the feed regularly, that first item will be different every time.
Let me click Save. From there, we want to select the first in the feed,
and you can see. It's already found the feed, and so now you just
click on the first one in the feed and click Save. Now, that is done. What you can do here is if you want to test this,
if you right click, you can say Run
this module and it will use one operation to do so, and we'll see if
it grabs the feed, which it has done, and it returns the information
that it's found. So it outputs it in
this structure and it shows you one bundle, which is one bundle of data. If we were grabbing multiples, then we'd have multiple bundles. And in here, you can see the ID, you can see the title, you can see the description, and you can see the URL. So it's actually working, and that is the first part. Okay, this is where
it gets interesting. We are going to
add another module by clicking that plus sign. From there, we want to
find Open AI chat EPT. You can search for
it in the apps. I'm just going to click it here. From here, we want
to add this one, which is it completes
a prompt or a chat. Okay, now it has appeared. Let me move it over a bit. First thing that you need to
do remember when you added your Open AI account and
you've got your your ID, your connection, your
code, your API key. That's what I meant. Go and grab that and add it there
and then click Save. That should I might
take a few seconds, and it will come back with
the screen that you're seeing here and it'll be ready for you to start your
chat completion. Next thing to do then is
you should get a list of all of the models that
are available to you. The one that we
want is 3.5 turbo. It's sufficient for our needs, and it's the cheapest
one, as well. So I'm clicking on 3.5 turbo
while that is grabbing it. I'm just going to
share another screen, which is to show you
I went to chatpt.com. And just to get an idea
of how many tokens I would get for my $10 that I created in my account
in my developer account, and you can see the
cost of them there. Won't labor this
too much, but you can see how very,
very cheap it is. If I come down here, I thought this was
interesting to share that Chat GPT four is 30 times more expensive than
the version 3.5 turbo. So if we want to keep it very, very cheap and it is
sufficient for our needs, then stick with 3.5 turbo. Right, number of tokens. I'm just going to put 200,
which should do the job. All we need to do now is
add a message in message. If I click the box here, I need in the role, I need developer stroke system. This is just to talk to
GPT as a system user. Now I need text to go
into Tell J. Chat GPT, what to do and what
response I want. When I first click that, I'll get this box appear
on the right. But please ignore
that just for now. I'm just grabbing
a little bit of data which I'm going to
input in here. Okay. This content that
I've put in here, this text is to tell Chat GPT what to do with the response
from the RSS speed. Again, you just copy
it word for word. If this is not making
a lot of sense to you, just copy this word for word, and it should still do the job. I will run through it quickly. What I am saying is take the
content below and create a tweet there's a maximum
of 240 characters, just to give Chat EPT a little bit of room
for its response, so it doesn't go to the actual
limit of the tweet limit. Include the title of the
post and a direct link. Include the direct link of
the URL in the content. Please remove all
additional query tags. The reason for that is that the URL that you
can see over here, which is coming
from the RSS feed, you can't see the
whole thing, but it's actually
really, really long. And the tweet what it
would do is it would click through to Google first and then it will go to the
URL that we don't want. We just want the URL. I've asked it to remove all the additional characters and just give us the content, the URL that goes
directly to the content. It works pretty well. And also, I've asked it to remove
any quotation marks at the beginning and at
the end of the response. Chat EPT sometimes tends to
coming from this system, at least, embed its response in quotes, which we don't want. And then another bit
that's important is, please include
relevant hash tag in the post and keep
within 240 characters. Again, I've reiterated that. You can have a play around
with that if it doesn't work. I've tried it. I mean,
I've tried this, and this actually works
for my feed really well. Now, can you believe
I almost forgot this. But where we've entered
the text, it says, take this content below, it doesn't know what
content to take. I'm going to add it now. We have to grab the
content from the RSS feed, and we're going to make
it work with title and then it'll return
under the description, and then under the long URL, which it is going
to work with and generate a simple URL and remove all those
other characters. So I hope that makes sense now. Take this content below
and create the tweet. Okay, based on the information that it finds from the RSS feed. Okay, then, and then we
click the save button. Now here I'm going to
let me run this module, and well, yeah, let me
just run the module, and you can see it's
grabbing the RSS feed, and then it is coming back with our output. So
let me have a look. I think I have to go messages, that's the actual message
that we want the result. Sorry about that.
That was the input. Now we're on the output and you can see what it's done and it's created created the link and it's given us
some hash tags. Great job. Next step is just
to send that to buffer. The final step then is
to connect buffers. We click to add another
module in here, we search for buffer. We need to create status update, and the same thing
will apply that you have to do your connection, you click Add, save there. And if your buffer is open in your browser or
you're logged in, then it should just
come up here and say, allow access, and it's
as simple as that. Once that is done,
coming back to mine, once that is done, it
does a connection, and it will show you or it will list the channels that you've
actually got available. So you could post this
content onto multiple ones, although you will incur an
error because obviously you need an image to
post on Instagram. But for us for now, with this simple scenario, we're just going to
be posting our text directly to this one, which is X with a long number is this is coming from buffer, and that's how it's labeled it. Right. What are we going
to send to buffer? We click that. This
now shows you. You can see it kind of starts flashing if I hover over them. We want the result of our chat GPT. So it's as simple as that. Click that little
green result button, and that's what we're
going to send a buffer, and we want to post
it immediately. We could extend this scenario, include to update this at scheduled times in case you
want to see what you're posting before it goes up
because we are doing this blind straight into
Twitter, and that's it. I will save that, and
there we have it. That is a little scenario. If I just come back to
the beginning down here, how often do we
want this to run? I'm going to say it
does this in minutes. Counseled out there
for 80 minutes. I'm going to save that. If
you look back down here, it says it's converted
that into 8 hours. When you set it down here,
this is now ready to go. When this is on. This will run every 8 hours and it's going
to call this RSS feed. And that's it. Job done. Let's just move on
to the final video, and I'll show you the output. But if you've got this far, well done, that's great result.
6. Checking the Results & Well Done!: Oh. That was quite a whirlwind run through to get
through that lot. If you are obviously already familiar with some
of those services, then you should
have got on Okay. Hope you did and didn't hit
too many snags along the way. This is my account. It's called AI Music, which is my speciality
on here on Skill Share. Check out some of
my other courses. I've only just put this live, so it doesn't have many
followers right now. All these entries have
good description, hash tags, and they'll all click through to their actual story. Okay, so what we'll do now
is we'll head to make.com. Right. I M, this is
finished scenario. We can set it to
run every 7 ounces, which I've done there or use this button to run it manually. So let's click that. Away it goes, you can see
it generating the RSS feed. Now it's moving into
generating the Chap GPT, doing its thing, and now it's
pushing it over to buffer. So we've got three ones, going on, which means that it should have
been successful. Great. All right, let's
have a little look. So this is the data
of the new record. That's the final result.
I'm going to pull up my original and do a refresh, and you can see that it's generated that
result in real time, looking, you know, pretty good. You can click through by
clicking anywhere on that tweet and it will follow through what Google has provided
in the link close that. Looking, really good. So good result,
working very nicely. I don't seem to have
many problems with it. Yeah, I hope you
managed to do the same. I think we're about done. One thing to mention
is that when you are creating content for
your social channels, you do need to keep
on top of them. I know, of course,
this process has been fully automated
into your feed, but there are no guarantees what will appear from the
RSS feed or in fact, what GPT will generate as well. I would say that this process
should really supplement your own content
and keep your feed updated regular with
this automatic content, but also craft your own content. I'm sure that would
probably help the Twitter algorithm as well. Great. Okay, well,
I really hope you enjoyed that and that you
will create your own feed. If you do, please put it in the project below.
I'd love to see it. And, of course, I
will follow you. If you do that, that would
be so cool. Great. Okay. Oh, Ann, just a quick plug here for my other AI music classes, which are there to help
content creators actually create media using AI music. Check that out if
you get a chance. Okay? That's it. Bye for now.
David Hughes, Producer | Creator | IT professional
