Hermes Agent [clears throat] is the most
popular AI agent in the world with
nearly 200,000 stars on GitHub. And
that's because it can do nearly
anything. But just like other agents,
Hermes often gets blocked when trying to
access the most valuable sites on the
internet. Luckily, there's a solution
that allows us to get qualified leads
for your business, find the best hires
for your company, track competitors'
prices and market changes, find better
job and career opportunities, and access
websites your agents generally get
blocked from. And it is surprisingly
simple to set up. In this video, I'm
going to show you how to set up Hermes
Agent, how to connect it to Apify so
that it can scrape anything, and my own
personal use case that allows me to get
an unfair advantage over any other
business in Europe. So first, pull up
the Hermes Agent GitHub repo. I'm going
to link that below the video. And then
scroll down to the quick install
section. There it is. Quick install.
This allows us to copy a single command
and install Hermes Agent on your
computer. So if you don't have any
powerful agent running locally on your
MacBook, there really is no excuse
because it's literally a single command.
So copy that and then open terminal,
paste that in, and just hit enter. And
this will install everything that Hermes
Agent needs to run on your computer.
Now, while that is running, we can
actually set up the Apify side of
things. It has over 40,000 different
tools, whether it's for TikTok, Google
Maps, Instagram, website content
crawler, e-commerce scraping tool,
Facebook posts, LinkedIn, Reddit, all of
the most difficult sites that normally,
with simple web browsing, these agents
have no hopes of accessing. With Apify,
they can do that super easily, which
then we can use Hermes Agent on to do
anything we want with that data. So
basically, Apify is one of the first
tools you want to connect to any AI
agent you build so that you make it
instantly 10,000 more powerful. And
that's exactly why I asked Apify to
sponsor this video, and they agreed. So
a huge shout out to them. Now, let's
check on our Hermes Agent. All right, so
the install has finished. So we can
select a the
As you can see, there is many different
providers. If you have a ChatGPT
subscription, you can use that. If you
want a great deal, MiniMax subscription
is amazing. Tokens per dollars, if you
have a lot of money, maybe you go with
Anthropic. Or if you want access any
model, you just go with OpenRouter. We
need to put an API key here. So, if you
don't have an API key here, just go to
OpenRouter and create one. As you can
see, I already have, so I'm going to do
K to keep it. And now we need to select
a model. Unfortunately, Fable got
banned, which is a tragic day in the
history of humanity. So, we're going to
have to stick with 4.8 fast. Keep the
current environment. We don't need to do
any messaging platforms right now, so we
can skip that. And we're done.
So, now if I do clear and Hermes, we
should be able to chat [clears throat]
with Hermes agent installed locally on
my MacBook with Opus 4.8 fast. Let's do
uh
Who are you? Classic. And there we go.
So, Hermes agent is responding, as you
can see, installing it was very easy.
Now, let's make it 10 times more
powerful by giving it the ability to
access any data on the web and scrape
literally any website, no matter how
restrictive or how messy the data is
from that website. And for that, we're
going to use the Apify MCP connectors.
And this is a new product from Apify
that allows us to use any of the 40,000
plus actors and connect them with things
like Superbase, GitHub, Notion, wherever
your data lives, right? So, the actor
can instantly scrape the data and save
it to your database. And then, we can do
whatever we want with that with Hermes
agent. And by the way, you can get
started completely for free with Apify.
So, click the first link below the video
and create an account. Then, go to top
right and click on go to console. This
will redirect you to the Apify console,
where you manage all your actors. You
can see the runs. You can create
automations, schedules, a lot of
different things. Now, the first we need
to do is go to left and we need to
choose which actor, aka scraper, we're
going to use to access the data we want.
So, the thing I'm going to be building
is a tool that allows us to find the
best hires for any role. So, usually
those people are on LinkedIn, so we're
going to use a LinkedIn actor, but this
could be used for many different ways.
Now, for LinkedIn, I found this specific
Apify actor, and this one's great
because it doesn't require any LinkedIn
login, no account cookies. It has lower
on-camera band risk because LinkedIn has
pretty restrictive uh terms of service.
So, do your own research before running
mass scraping operations. It searches by
job title, location, industry, and
returns richer profile data for Hermes
to analyze. So, I'm going to be using
this specific Apify actor. And by the
way, everything from this video is going
to be linked in the second link below
the video, completely for free in a
bundle. All of the setup, the MCP
config, the prompts, the list of Apify
actors, everything you need to follow
this video is included in the second
link below the video. So, just click on
that and grab that. It's going to be
completely free. But, let's say you want
to use a completely different Apify
actor. Well, inside of the Apify
console, go to left, click on actors,
and then top right you can see go to
store, and here you can see all of the
available 40,000 plus different actors.
And they're split into different
categories for lead generation, rising
stars, popular actors from the
community, and many, many more. So,
whether it's YouTube, Reddit, LinkedIn,
Twitter, or anything else, you can
probably find an Apify actor for that
thing. Okay, so next we need to set up
this Apify actor and to do a test run to
see whether it works or not. So, inside
of the profile scraper mode, I'm going
to click on full. You can also do full
plus email search, up to you. Search
query. So, this is what you want to
search. So, I'm going to do AI developer
or software engineer or programmer,
something like that. We can keep this
maximum number of profiles to 20.
Location filters, I'm going to add some
local things because we're building an
office in Katowice. If you want to join
us, by the way, there's going to be a
link below the video. Then, go back to
the top, and in the top right you can
see save as a new task. Click on that,
name it something like Katowice AI
developer, whatever, just a descriptive
name. And then click on continue. And on
the left, by the way you can see we got
redirected into saved tasks, where you
can see these like kind of presets for
these actors, so you don't have to
configure them every single time. You
can just make sure you have the query
here, software developer.
This is most important, otherwise you'll
get an empty run. So, I filled it in
again, and we should be good. So, click
save and start, top right.
And it should start a run
with this actor, and it's looking for
software developers in this area, and
it's already finding results. Look at
this. It's amazing how fast this is. So,
at the top you can see the spend, and we
can also go to the top left runs to see
all your previous runs, and the status
of this run, whether it's finished or
it's still running. If you want to see
how many credits you have, go to the
bottom and click on usage, and you can
see the dollar you spent, and the amount
of dollars you have left. If you want to
upgrade, just click on upgrade and
select the plan that makes the sense for
you, depending on how much how many
leads you want, how much scraping you
want, how many Apify actors you want to
run. Okay, there we are. So, we're
finished running, duration less than 2
minutes. And when you click on the
specific run, you can see actor
succeeded with 20 results in the data
set. Now, to view your data set, go to
storage, and here you can either export
it, or you can also view it in a new
tab. So, you see the raw JSON, and you
can see this is clear structured data.
This is like the stuff that AI agents
dream of, structured JSON that's easy to
understand, that's easy to work with.
So, yeah, this is perfect for our AI
agents. So, next, what we need to do is
we need to wire this up to Superbase
through the new MCP connectors, so that
anytime we run this Apify actor, and
again, we can do that on a schedule,
because Apify makes it very easy inside
of their console to set all of this up.
It goes to Superbase database, and then
we can access that through our Hermes
agent to analyze these leads, to contact
them, and just do all the other
operations you would want to a data set
of contacts, whether it's leads, whether
it's potential clients, whether it's
guests for a podcast. Again, get
creative. You can absolutely grow your
business or improve your life if you
implement scraping in the right way. For
example, every business wants more
leads. Every business wants more
customers. If you set up your Apify
actor in the right way, you can
absolutely find these people and you can
set up a Hermes workflow to
automatically reach out to them. So,
before we can implement Hermes agent so
that it takes the data set created with
our run and it can analyze the leads,
qualify them, reach out to them,
whatever you want to do, we need to set
up the Supabase integration so that
everything that happens with our actor,
all of the leads that it generates, all
of the contacts that it scrapes, it
saves into Supabase. So, just go into
Google and type in Supabase. The free
account is more than enough. Just click
on start your project. So, since I
already used up my free quota, I need to
pay $10 a month, but if you're starting
on Supabase, you can have two projects
completely for free. So, just go with
the free Supabase account. You really do
not need to pay for this one. Okay, so 2
minutes later, we're up. Go to the left
SQL editor, click on plus new snippet,
and I have a simple table generation for
leads. And again, all of this is in the
free bundle, second link below the
video. All the prompts, all the SQL
presets, Apify actors, everything. It's
going to be in that bundle. So, you can
see this was successful, no rows
returned. So, now if we go to table
editor, we can see that we have this uh
table leads.
We might need to set up uh endless
policies for that in a sec, but uh I
think this could be enough for now. So,
inside of Apify, you might be on runs or
saved tasks. You need to go to settings
at the bottom. Settings, go to API and
integrations here. Boom. Scroll all the
way down and MCP connectors, okay? This
is the section we're looking for. Click
on add a connector. So, in the name, put
something like Supabase Apify hiring
demo. And here, we're going to actually
extend the URL by adding our project
reference there. Go back to Supabase
project overview. And this is your
Supabase URL right here. So, just copy
that and put that after equals. Next, we
need an API key, and for that, it's It's
a project one. It should be a personal
access token. So, in Superbase, go to
{slash} dashboard {slash} account
{slash} tokens. Access tokens, generate
a new token. I can put it at expiry like
1 day or even 1 hour. I'm going to name
the token Apify MCP demo.
Expiry, just do whatever you want, or
you can do never, that's risky.
And
click on generate token. Now again, do
not share this with anybody, this can
access your entire account, so treat
this with respect. Obviously, I'm going
to delete this token before uploading
this video. So here, copy the token
because as it says, you will not be able
to see it again. Go back into Apify and
paste it in here. Click save, and MCP
connector should appear right here at
the bottom in the list of MCP
connectors. All right, so we're going to
use this Apify actor, universal MCP
connector. So, click try for free. This
is going to take you into the Apify
console. As you can see, this actor
requires the dataset ID. So, let's go
back to the runs, let's click on the
last run, the storage again, and here
we'll see the dataset ID. Copy that and
give it to him.
Wait, it's loading. Boom, click that.
Nice. The user query here, we need to
replace that with something like this.
Don't overthink it. The language model,
keep it to Deep Scrape before flash,
that's perfect. Max agent steps, put it
to like 40, 100 is unnecessary. And for
the MCP connector, select this one,
Superbase Apify hiring demo. And
actually, let's add one more thing to
the query. If a row already exists with
the same profile URL, skip it. So then,
scroll back to the top and click on save
and start.
This will start a new run,
and hopefully, it'll get uploaded into
our Superbase. So, let's go here, let's
go table editor, leads, and we should
start this getting populated in a second
by Apify itself, so it should be all
good. So here, we can refresh this to
see if it's populated yet.
All right, not yet. Here, I'm going to
switch the query to use the dataset, and
this is the name of our dataset from
before, so you can go to runs and click
on that run,
and again, storage,
and here you can see the data set ID.
So, that's that name.
We're going to test the first five items
only. Here is the structure
of how to map it into Superbase.
And for the model,
let's use something a bit more
competent. I'm just going to go with
Sonnet 4.6. It's like a mid-level model.
Not too expensive, but not a super small
model either. So, here I'm saying use
data set and that's the data set name.
You can access that by going to that
run.
Uh runs. You know, you click on the run
we did before, storage, copy that, boom.
So, I extended the prompt slightly,
included the Superbase name, specific
name of the table. I'm going to save
that.
Model we're using Sonnet, agent steps
40, select the MCP connector here. And
we can start the run. Let's see if this
actor works. Okay, so the actor is
working. We can click on logs to see
what is happening.
It is connecting to the endpoint.
Doing some LangChain tools.
When it's converting the MCP tool to
LangChain tool, okay.
It spent 1 cent, so it should be doing
something.
But we got an error. So, let's go to the
logs. Okay, so it seems like uh
the run succeeded. My API key expired
because not the API key, the personal
access token expired because I put 1
hour on it. But after I fixed that, we
have a successful run here. So, if we go
to Superbase, into leads table, there we
go. We have the leads here.
Amazing. So, the first five leads have
been scraped and saved into Superbase.
This is massive progress. Now, we can
try accessing this with Hermes agent
so that he can qualify these leads,
analyze them, and then based on that we
could set up like follow-up workflows,
cold outreach, email, voice, even like
with Vapi, you know, 11 Labs. If you
wanted to call these people. Yeah, this
is a huge step. Now, the scraping is
being saved to Superbase. So, this
unlocks a whole new world of
possibilities. So, the next thing I'm
I'm to do is I'm going to screenshot
this, actually.
Copy, and I'm going to go to Hermes. I'm
going to say, "Just hey." And I'm going
to tell Hermes to help me set this up.
Right? So, I'm going to give it a
screenshot. Boom, control V. Okay, so
your task is to help me set up access to
my Superbase. This is for
Apify scraping. It finds the leads and
saves them into the Superbase table.
Tell me what you need from me to set
this up. Be very concise. So, you can
literally tell Hermes agent to help you
set this up, to tell you, to guide you
what it needs from you.
Tell us, "Okay, it needs a service role
key and a project URL."
So, say, "First, find the
proper place in Hermes agent
where to store ENV keys, ENV vars, and
secrets." Okay, there's a skill for
this, actually, Hermes secrets ENV.
Found it. Okay. So, it needs these
things, the Superbase URL.
So, again, we can go to project overview
and just give it that. Boom.
Here is the project URL.
Or the Superbase URL. Boom.
What do you need next?
Probably going to ask for the service
role key. But, we saved the URL into the
ENV.
And yeah, it needs a service role key.
So, we go to the left, inside of
Superbase project settings, API keys,
and it's the secret key right here.
Again, do not share this with anybody.
I'm going to rotate this key before
uploading the video. Copy.
Boom.
It's Okay, we need to allow that. Here
is the service role key. Save it.
Boom. Paste that in. Allow Hermes to
save it to the ENV. And I think that
should be
all. It should That should be everything
it requires to hit the database.
Okay, I have the full schema.
Leads table. Okay, Apify ready. Just
check the contents of the leads table
and let me know if you see the people in
there.
I don't think it needs the actor ID.
Yes, it can see them. Amazing. So now,
here's what happened. We ran Apify with
this LinkedIn Scraper actor
to generate these leads. Then, we used
this another Apify actor. I I told you
guys they have actors for basically
everything. This one is called Universal
MCP Collector, and it connected to
Superbase through the MCP.
And then, now we gave the essential We
we gave two things, the Superbase
project URL and the service role key to
our Hermes agent. It can now work with
our Superbase. So now, basically, what
remains is automating both steps, both
the Hermes agent analysis and putting
the actor on a schedule. So first, I'm
going to do a test analysis to see how
Hermes agent does this by sending it the
prompt. So here's the full prompt.
Connect to my Superbase project and read
the public leads table. Score each
candidate from zero to 100 for fit for
my Katowice office. Some basic uh
fix to prioritize and how to update the
table. So that's the full prompt I just
sent to Hermes agent. Let's check how
it's doing.
I've all records, all five scored and
written back to public leads. Lead score
rating, reasoning, fit tags, and rated
all.
And we can check actually
inside of Superbase, back to table
editor, leads to see if the scoring is
here. There is lead score. So one guy is
Okay, these two are promising fits. This
is very useful already. So instead of me
wasting time interviewing all these
people or reaching out to them, Hermes
agent figured out these two people are
worth of interviewing, and these three
are not. Like guys, do you see how
insane this is?
Apify plus Hermes agent is really
incredible combo. And again, this is not
just for hiring. This is not only for
finding employees. This could be for
finding leads, B2B, enterprise
customers, sales. This could be for
finding dates, this could be for finding
friends in a new city, this could be for
finding
great restaurants, gyms, for checking
actual like if you want to, you know,
co-working place or cafe, you don't just
want to look at the reviews, you want to
scrape all of the comments, all the
reviews to see is there fast internet,
is there good seating, all this stuff.
There's so many things this can enhance
your life, grow your business faster,
you just need to get creative. Now you
have all of the steps, so sit down and
think where can I implement scraping
with Apify and Hermes
so that
my life is advancing faster. What
problems can I solve with just a simple
Apify actor and connecting that through
Superbase or GitHub or Notion into
Hermes where the Hermes does the
analysis, does the reach out, does the
qualification?
Really sit down, spend a few minutes
thinking about this because I guarantee
you you can find multiple different use
cases in your life or your business how
you can make this immediately useful.
Now as I promised, let me show you how
to fully automate this both on the Apify
side and on the Hermes side, starting
with Hermes.
I'm going to send this prompt. Every day
read the public.leads table in Superbase
and find all rows where lead score is
null.
Score each new lead from 0 to 100 for
fit for David's Katowice office,
prioritizing software engineering
ability, AI automation experience,
seniority, local relevance, blah blah
blah.
And evidence of shipping real projects.
Write back lead score, rating reasoning,
fit tags, and rated at for each new row,
then send David a short top five digest
with whom to contact first and why. This
already will save my team
dozens of hours per week, okay?
Otherwise humans have to do it. Like
when you're getting hundreds of
applicants like we are getting,
you need to go through them, right? And
if you don't use an agent for that, a
human has to do that. So Hermes will
probably create a cron job for this.
Let's scroll down, let's see what it
did.
It's asking for command.
Security pipe, we're going to allow
that. Script works.
So now it's preparing the cron job to
automate this daily. Cron create
Katowice lead scorer and done.
This is the name of the cron job. Here's
the ID, schedule 8:00 a.m. every day,
and it will deliver here. Obviously, if
you want to connect Discord, Telegram,
or Slack, you can easily do that. In
Hermes agent, I showed that in many
other videos. So, feel free to watch
these after this one. But now, let's go
into Apify. On the left, click on
schedules. And now we will automate the
scraping. So, actually first, make sure
you did the saved task, right? As I
showed you before, uh this is how you
can create this kind of preset with the
query. So, I just did software
developer.
And you can customize this way more.
Apify is highly, highly customizable.
So, this is going to be the saved task
that runs on a schedule. So, when you go
to schedule, click on create new. We can
put this actually hourly.
Instead of every hour, let's do custom.
Let's do something like every 6 hours.
That should be good enough. So, that's
four times a day. Actually, I'm going to
rename it to SWE search 6 hours.
Automation that runs every 6 hours to
scrape for more software engineers.
Boom, save that. Then at the bottom,
click on save and enable. And then
scroll down below and select the actors,
right? So, we don't have any actors or
tasks. We're going to select the
scheduled task.
Boom.
Here we have it.
Education developer.
Here we have the one we created. So,
select that and click save.
Save and enable.
Enable the schedule.
And it's saved. So, now if we go to
here, we can see that the SWE search 6
hours is enabled. Notifications are
enabled. There's the frequency
described. And we're going to see when
the next run is happening.
And just like that, we just automated
the scraping. We automated the
qualifying and the analysis.
And this alone, if you put this as a
business, would put you ahead of 19
I mean, like 99% of companies, easily.
Like how many companies you know that
have automated their recruiting
process, their head hunting, their their
hiring process
and their analysis, you know, big part
of their HR. This is just automated.
Some of this is programmatic, some of
this is agentic, but this is going to be
costing a few dollars per week
and it's going to save thousands of
dollars per week in terms of salaries.
If I had to hire another HR person or a
recruiter or a headhunter,
like it's insane what you can do with
some code and AI agents. So, yeah. This
is the power of Apify plus Hermes. Many
people are underestimating
underestimating what's possible with
scraping and uh with Apify, you can
scrape basically any website you can
imagine, even the most difficult and
restricted websites. Obviously,
research the terms and conditions if you
want to do like massive scraping
operations, but
it just works. Like it's easy to set up.
I showed you all the steps in this
video. So, if you only watched it, go
ahead and go through it again and
actually follow all the steps and
implement it into your business right
now.
Whether it's for hiring or for getting
more leads or for anything else, doesn't
matter. And again, to get started with
Apify, click the first link below the
video and create an account. And to get
all of the resources, prompts, configs,
SQLs, everything I showed you in this
video, click the second link below the
video and grab this free bundle. It's
completely free. So, with that being
said, I wish you guys a wonderful
productive week. See you.