YT YouTube2Text

YouTube2Text → Transcripts → make f*ck you money with claude fable 5.2 + free github repos

make f*ck you money with claude fable 5.2 + free github repos — Transcript

by Corey Ganim · 5,118 words · 695 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Claude Fable 5.2 has been out for less
0:03than two days, and I've already spent at
0:05least 10 hours using it, and I'm about
0:07to show you exactly how we can use it to
0:09build an entire business for us from
0:11scratch from a free, completely
0:14open-source GitHub repo. So, we are
0:16going to literally have Fable 5.2 build
0:19us a business in a box. And I'm going to
0:21use this exact prompt to do it. So, by
0:23the end of this episode, we are going to
0:26have a business idea. We're going to
0:29have Claude actually go out and find
0:30three different GitHub repos. We are
0:33going to um have Claude build us an
0:35offer on top of those three repos. We're
0:37going to choose our favorite one. Then
0:38Claude is going to go find 50 leads that
0:42we can cold email or cold call. It's
0:43going to build five different static ads
0:46that we can run on Meta. And then it's
0:47going to build us two different UGC
0:49videos that we can run on Instagram or
0:51Tik Tok. So literally, we are going to
0:53have a turnkey business here in about 20
0:56to 30 minutes time. So, this is the
0:59prompt that I'm giving to Claude. I'm
1:01going to include it for free at the link
1:03in the description. We're going to add
1:04this prompt to the classroom of our
1:06completely free uh AI operator hub
1:09school community. So, if you join that
1:11community again, which is completely
1:12free, you're going to get access to this
1:14prompt along with the full playbook uh
1:16of everything that I'm going over in
1:18this video. So, again, we're going to
1:19let Claude do its work here. I'm going
1:21to talk through the prompt in detail so
1:23that you know exactly what it's doing.
1:26So, this is us basically telling Claude
1:29like, "Hey, Fable 5.2 just came out. I
1:31want to stretch it to the max of its
1:32abilities. I want you to build me a full
1:34business from scratch." So, the goal of
1:36this session is for you, Claude, to go
1:39out and find me three open- source
1:41GitHub repos. So, I'm literally running
1:43this from scratch. I have no idea what
1:44it's going to find us. Uh, come up with
1:46an offer that we can build on top of
1:48each repo. build an ICP, so an ideal
1:51client profile so that Claude can
1:53basically tell us who's going to be more
1:54likely to pay us for this offer, a price
1:57point for each offer, a full brief that
2:00explains our offer in detail, and
2:02include a style guide that we can use to
2:04actually generate the ads and um the
2:07emails and the cold the scripts and
2:09everything that we're going to use to
2:10actually promote this offer. A full
2:12script that can be read by the AI avatar
2:14for the UGC video. use a tool called TEG
2:18to find 50 businesses who fit our ICP,
2:21including their email addresses and
2:22phone number so that we can cold email
2:24and cold call them and then generate
2:26five variations of static ads that we
2:28can run on Meta and two variations of
2:30UGC videos that we can post to Tik Tok
2:33and Instagram. So, this is pretty much
2:36our business in a box and Claude is
2:37going to build this for us. Now, I even
2:39added a few different caveats. So, like,
2:41hey, this is a brand new business. We
2:42have no website. We have no meta ads
2:44yet. So for the static ads, I'm actually
2:47using a skill that is a free open-
2:50source skill from graph.com. So if you
2:53go to graph.com, I have no affiliation
2:55with this company. Uh it's run by Cody
2:57Schneider. He's someone who I really
2:58enjoy his content on X. They kind of
3:01brand themselves as like the marketing
3:03engineering cloud. If you go to
3:04graph.com and scroll down to the bottom,
3:06they have these skills and all I did was
3:08copy setup prompt, handed to Claude.
3:10Claude installed the skills. The
3:12specific ones that I used are UGC video
3:14and static ads. And I added like $15 in
3:18credits to my Graph account to for
3:21Claude to be able to run those ads. And
3:23typically, you have to give Graph or
3:26sorry, you have to give Claude a website
3:28URL or a meta ads link and it's going to
3:30go and copy the successful ads of those
3:32other companies or build the ads from
3:34your website. But again, because this is
3:37a brand new offer, we're basically
3:38saying like, look, skip that step. use
3:40GPT image 2 to simply build these ads
3:43from scratch and then um we gave it a
3:47few checkpoints that we want it to
3:49follow. So, you know, let me know once
3:51you found the three repos and created
3:52the offer. Then let me know once you've
3:54created the ICP doc so that I can
3:56approve. Then let me know once you
3:57pulled five businesses from Tre so I can
4:00make sure we're looking in the right
4:01place. And then let me know when you
4:02built the ads. So for checkpoint number
4:05one, it already found us the three
4:08repos. uh including the number of stars
4:11and the actual number of commits they
4:13have. So the first one that it found is
4:15called Tasty Igniter which is a
4:17restaurant online ordering and table
4:19reservations tool which sounds super
4:21interesting. The second one is called
4:23Dog Raw self-hosted voice AI platform.
4:25It's an alternative to Vappy or Retail
4:28and then Chat Woot which is an omni
4:30channel inbox for chat email and social
4:32DMs. So what is the offer that Claude is
4:34actually recommending that we build on
4:36top of each of these? So for Tasty
4:38Igniter, it's saying that we should
4:40build a commissionfree direct ordering
4:42site. So I know a lot of restaurants
4:44hate Door Dash and they hate Uber Eatats
4:46because Door Dash and Uber Eats take
4:48such a big cut of the bill, like 30% in
4:51some cases. So it's basically saying
4:53like, look, let's build a ordering site
4:56that gets the restaurant that 30%
4:58feedback and we can monetize via like a
5:00subscription. We could use Dogra to
5:02build a 247 AI phone receptionist, which
5:05I know is a super popular offer that I
5:07see a lot of people talking about
5:08online. And then with Chat Woot, it says
5:11create one inbox plus an AI first
5:13responder.
5:14And then it even says who's the buyer
5:16for each of these offers. So for Tasty
5:18Igniter, target independent restaurant
5:20owners for dog home service contractors,
5:22which I feel like this AI phone
5:25receptionist and targeting these home
5:26services like that's super overplayed.
5:29Like everybody's trying to do that. So,
5:30I already don't like this second one
5:32here. And then for chat woot, it's
5:34saying target med spas and salons. So, I
5:36actually like number one or number
5:38three. I think for the sake of this for
5:42the for the sake of this exercise, let's
5:44go with number three because I feel like
5:46that is a huge need um for these com.
5:49Actually, you know what? Let's go with
5:50number one. I'm changing my my mind mid
5:53mid-sentence because I actually have not
5:56targeted restaurants in any of my offers
5:58in the past. I think it's super
5:59interesting and I think it's a very sexy
6:01offer because like I said a lot of
6:03restaurants they Door Dash and and Uber
6:05Eats is a necessary evil, but if there
6:08was an opportunity for them just to pay
6:09like a $149 a month flat fee, I think
6:12that'd be a lot more enticing. And
6:15Claude in this case is saying it's
6:17essentially commission free direct
6:18ordering. The offer was would be we
6:20build the restaurant a branded ordering
6:22site in seven days with their full menu
6:25loaded, card payments connected, QR
6:27codes for tables and bags, and the
6:29Google order online button pointed at
6:31their own site. They pay a flat fee and
6:340% commission. Why they pay is because
6:37delivery apps take roughly 15 to 30% per
6:39order. A restaurant doing 8K a month
6:41through the apps pays about 2 grand in
6:43commission. So moving a third of those
6:45orders direct roughly covers our fee
6:47four times over. So to me that's a very
6:49compelling offer. So I really like
6:52number one and it's saying two things to
6:54settle before later steps for UGC
6:56scripts. The business has no customers
6:57yet. So I will write the avatar as a
6:59spokesperson talking to owners not as a
7:01fake customer testimonial which I agree.
7:04Uh let's go with tasty Igniter. So we
7:06are literally just going to click the
7:07button that says you know build the
7:09Tasty Igniter offer. We are going to let
7:11Claude go to work. And while it's doing
7:14that, let's pull up Tasty Igniter
7:19on GitHub and just kind of take a look
7:21at this repo. So, it's got 3.8K stars.
7:25So, you know, it's it's popular, but
7:27it's not like hugely popular. This is
7:29not something that like everybody knows
7:31about. Uh, open- source online ordering,
7:34table reservation, and management system
7:35for restaurants. They've even got a
7:37website, tastyigniter.com.
7:40Very cool. So, again, I I'm a fan of
7:42this. I'm excited about this offer. I
7:45think this is something that we can
7:46build relatively easily with Fable and
7:50Wild Claude is going to work here on
7:53building out the next part of the ICP.
7:55Wow, it's already done. So, it's reached
7:57checkpoint number two, which is the ICP.
8:00So, it's telling us the buyer for this
8:03offer would be the owner operator of an
8:06independent familyrun Mexican restaurant
8:08or Takaria in a major Texas metro with
8:11loyal regulars who already sells through
8:13delivery apps and has no ordering system
8:15of their own. So, interesting because
8:17Claude Fable went very specific with
8:19this ICP. Personally, I wouldn't have
8:23thought to target just Mexican
8:25restaurants. I would have thought to say
8:26like, hey, just like owner operator of
8:29independent restaurants, period. But it
8:31went a layer deeper and said, let's
8:32target Mexican restaurants and let's
8:35target them in a major Texas metro
8:37because there's probably a lot of them.
8:39It even said, I narrowed to one cuisine
8:41and one state on purpose. It makes the
8:43ads in cold call specific and matches
8:45the Latina kitchen avatar, which is one
8:49of the avatars that we have available to
8:51us for the UGC skill that we got through
8:53graft. So that's actually a very smart
8:56call on the on the part of Fable 5.2
8:59and it keeps the calling in one time
9:01zone. The offer works for any takeout
9:03heavy restaurant. So widening later is
9:05easy. I completely agree. So what it's
9:07saying now is we've got all these
9:09different criteria that our ICP needs to
9:11meet. So again, independent Mexican
9:13restaurant, one to three locations in a
9:15major Texas metro, proof of regulars as
9:18evidenced by a 4.2 two star rating or
9:21more and 150 or more Google reviews.
9:24Roughly 500k to $2 million revenue per
9:27location. And again, that's something
9:28that TRE can verify for us or at least
9:31look into. Um, we can verify whether
9:34they are listed on Door Dash, Uber Eats,
9:36or GrubHub. We can see if they have a
9:38website, but the ordering on the website
9:40goes directly to an app or is phone
9:42only. And make sure they have a public
9:45phone and a findable email. Now again,
9:47it's even telling us how do we check it.
9:49So TEG, which is a tool that I love, uh
9:52I'll even show you guys what that looks
9:53like. And again, not affiliated with TEG
9:56in any way, but it is the open router of
9:58AI subscriptions. So instead of me
10:01having to go and subscribe to like
10:04Simrush or Apollo or, you know, any of
10:06these individual subscription tools, I
10:09can pay per usage via Tre and Tre is
10:12just hooked up to Claude. So Claude can
10:14do all of this work for us. So, it is
10:16saying it's going to use the Google Maps
10:18connector in Tre. Uh, it's going to use
10:20Google Maps city search to find the
10:21location, Google Maps rating and review
10:23count to find proof of regulars because
10:25Tre has full Google Maps API access. And
10:28we're going to exclude any chains and
10:30franchises. The specific person we're
10:32targeting is the owner operator, runs
10:35the business from a phone, uses
10:36Facebook, Instagram, and WhatsApp. Um,
10:38often bilingual. We open in English and
10:41offer We open in English and offer
10:44Spanish. Okay. Interesting. So,
10:45obviously with AI, we can build the site
10:47to be fully in Spanish if we need to.
10:50Price ceiling is about 200 a month. No
10:52contract. They decide within a week. So,
10:55it even goes into like what hurts, like
10:57what is their actual pain point. Um, so
10:59I'm not going to get too deep into this
11:01now, but it tells us like what makes
11:03them act now, what objections are they
11:05likely to have, when could we where, and
11:07when could we reach them? So, it even
11:09says like cold call them on a Tuesday to
11:11Thursday between 2 and 4 p.m. because
11:14that's between lunch and dinner service.
11:16Cold email to the owner or the general
11:18inbox with their own commission math in
11:21the first line. Um, so this in my mind,
11:24like I think this is really strong so
11:25far. Qualifying rule for the lead list.
11:28The return is obvious when a restaurant
11:30does about 4K a month or more through
11:32the apps. at a 25% commission, moving
11:34about $600 per month of orders direct
11:38covers the $149 fee. Obviously, we can't
11:41see wrap app revenue from the outside.
11:43So, the review count is what we're using
11:45as a standin for determining how much
11:48revenue they're actually doing through
11:49the apps. So, now it's asking, do you
11:51agree with this ICP including Texas
11:54Mexican restaurants as the first niche?
11:56I'm going to click approve ICP.
11:59You notice how it gave us an option to
12:00say like approve ICP or the other option
12:03was all cuisines in Texas. So if we
12:06wanted to go ahead and widen our search
12:08a little bit, we could just say like,
12:09hey, targeting Texas is still fine, but
12:12let's go with all types of cuisines, not
12:15just Mexican restaurants. Personally, I
12:17really like having a very specific ICP.
12:19I think the more niche you can get, the
12:20better. And if we can prove this out in
12:23the, you know, Mexican food/takaria,
12:26uh, niche, we can always expand that
12:27later. So, right now, it's going to go
12:30ahead and write the offer brief, the
12:32style guide, and the style guide is what
12:34we're going to use to generate the
12:36static ads, and it's going to go ahead
12:37and write the script for the UGC ad as
12:40well. So, again, I have no idea like
12:42what this is going to look like. This is
12:44my first time testing this particular
12:46workflow with Fable. That's the point of
12:48this episode is to just see what CA
12:50Fable is capable of. See if it's even
12:52possible for it to build us a business
12:54in a box. My gut tells me is it's going
12:56to do a really good job, but you will be
12:59you and the audience can kind of be the
13:01judge of that here when we get to the
13:02end of the episode and see the full
13:04output. So, I'm going to pause and come
13:06back once this is ready to go and we'll
13:09see where we go next. So, it's not quite
13:11done yet, but I wanted to point out the
13:13fact, and this is one of the reasons why
13:14I love Trey, is it used Tre. It said it
13:17already returned 40 Mexican restaurants
13:20across Houston and San Antonio with
13:22their phones, their ratings, and their
13:23websites. Then, it's going to check each
13:26restaurant for an ordering system, then
13:28look up their emails, and then draft the
13:30brief. And it used any mailfinder via
13:34TRG to uh look up the email addresses.
13:37So again, any MailFinder is a tool that
13:39you normally you'd pay a subscription
13:40for. I think it's like 20, 30, 50 bucks
13:42a month, something like that. And
13:45because we're using it through TRG, we
13:46pay per use. So it says it cost 2 cents
13:50per domain where an email is found. So
13:52about 10 cents for this sample of five
13:55and about a dollar for the full list of
13:5850 uh businesses when we have it go out
14:00and find the whole list. So again, this
14:02is why I like using like an open router
14:05for LLM inference or a TRG for just
14:09usage based tool use because I only have
14:12to pay a dollar instead of paying like
14:14$50 a month for something that I'm only
14:15going to use one time. So I just wanted
14:17to, you know, unpause, make that point.
14:19Right now it's still writing the offer
14:21brief, the style guide, and the scripts
14:23for the UGC. So I'll be back when that's
14:25done. All right, so it took only about
14:27five or six minutes, but Claude is back.
14:30So it we are now at checkpoint three,
14:32right? So checkpoint three is it went
14:34out and pulled five sample leads and it
14:36wants me to approve the leads first
14:38before it goes and pulls the other 45
14:40for a total of 50. And it also went
14:43ahead and created the brief, which I'll
14:45show you guys in a second, which lays
14:47out our ICP, the full offer, as well as
14:50the creative and style guidelines that
14:53Claude is going to follow to produce the
14:56ads. Now, why is that important? Well,
14:58if we don't give it style guidelines or
15:00any sort of like guidelines to stay
15:02within when creating ads or any
15:04creative, it's just going to give us a
15:06bunch of generic crap. So, what it said
15:09here is it wrote the brief and it used
15:11the working name Mesa Direct as like our
15:14business, which I think is interesting.
15:15I I think that's perfectly fine. It
15:17covers the offer, the pricing, and the
15:19owner's math, the ICP and the ad style
15:21guide, plus the five ad concepts and
15:23both UGC scripts. So, I'll show you guys
15:25what the brief looks like here in a
15:26second. Here are the five sample leads
15:29that it found. So, for example, one of
15:30them, Font Fonda Santa Rosa is the name
15:33of the restaurant. It's in Houston.
15:35Here's their phone number. Here's the
15:37email address. David Reyes at fondant
15:39rosa.com. Most likely the owner. 4.6
15:43Google uh 4.6 stars is their average
15:46Google review out of 1618 reviews. So,
15:49this is definitely a restaurant that is
15:51doing a lot of volume through Door Dash
15:54where a solution like ours would most
15:56likely save them a lot of money. So, it
15:57says for ordering today, they've got a
15:59Door Dash linked on a Wix site. So,
16:02that's a perfect perfect lead for us and
16:04somebody that very well might be open to
16:06our solution. Obviously, we've got four
16:08other leads here, very similar. All of
16:11them have, you know, the smallest one
16:13has 883 Google reviews. um Door Dash
16:16storefront, Door Dash linked on Wix
16:19site. So again, these are like really,
16:20really good leads for us for this
16:21particular offer. And it just gave us
16:24some notes about these particular people
16:26in these roles. What happens when you
16:27approve? It's going to generate 45 more
16:30leads with the same method across
16:32Houston, San Antonio, Dallas, Fort
16:33Worth, and Austin. It says the cost for
16:36finding those leads is roughly$1 to $2.
16:39It's going to next it's going to
16:40generate five static ads at about three
16:43and a half graft credits each. So like I
16:46said I put about $15 into graphs
16:48platform for credits. $15 I think is
16:511,500 credits. So it literally is going
16:53to cost us what is that 3.5 cents um or
16:57sorry 3.5 cents per ad. So to create
17:00five static ads it's going to cost us
17:02like 20 cents or something like that.
17:04And then for the UGC videos, it's going
17:06to cost us about $4.57 per video. And
17:09again, we're going to have it create uh
17:12two different UGC videos. So before we
17:14have it do that, I'm going to show you
17:17what it created for us. So it's working
17:19inside this Fable test folder that I
17:22created on my desktop. So inside that
17:24that folder, it created the subfolder
17:25called clients and in this case is Mesa
17:28Direct. And then inside the Mesa Direct
17:30folder, it created the brief, which I'll
17:32show you in a second. It even created a
17:34subfolder including all the leads in a
17:37CSV and the UGC scripts in a in two txt
17:43files as well. So like I'm already
17:45really impressed at just how proactive
17:47Fable was and just like organizing this
17:49data and putting it in a place where it
17:51can be easily accessible. Now showing
17:52you what this brief looks like. So it's
17:54going to open it up in cursor just
17:56because that's what I use as my default
17:58um like markdown reader.
18:01So this is the offer brief. It even says
18:03like your working brand name built on
18:05Tasty Igniter, the offer in one
18:06sentence, what the restaurant gets. So
18:09this is where it's laying out the offer
18:10in detail, which to me is very
18:13compelling, the price, the math that we
18:16show an owner, right? So this is where
18:18we break down like look, you're
18:19currently paying uh two grand a month if
18:22you stay on the apps. If 30% of your
18:25orders move direct, then you're going to
18:27save x amount of dollars, right? not
18:29only are you going to save money on the
18:32commission, but on the processing fees
18:33as well. So, that's where we kind of lay
18:35out the math that we show an owner, our
18:37economics, so like how much money we
18:39actually make, our ICP profile. Um, it's
18:43also going to lay out the messaging that
18:44we're using. So, even down to the
18:47objections that we're likely to get, and
18:49then the ad style guide. So, again, this
18:51is really important for generating
18:53static ads that are actually compelling
18:55and don't just look like some AI
18:57garbage. So Claude Fable 5.2 went ahead
19:00and created like look these are the
19:02exact colors we're going to use. These
19:03are the hex codes that we're going to
19:05use. This is when to use them. So like
19:07we're going to use the mesa cream color
19:09um for the main background. We're going
19:11to use chili red as the headline accent.
19:14Right? So it went super deep into detail
19:16layout imagery copy rules the five
19:19static ads concepts. So like the angle
19:22on ad number one is going to be
19:24commission. The angle on ad number two
19:26is going to be their regulars. The angle
19:27on ad number three is going to be their
19:28customer list. So on so forth. So it
19:31crushed it with this. And then lastly,
19:33the UGC video scripts. Here's variation
19:35A. Here's variation B. And then outreach
19:38compliance. So I'm again super impressed
19:41with what it's given us so far. Fable
19:43again is kind of just walking us through
19:45this step by step. It's like, all right,
19:46next step. Checkpoint three. The offer
19:48brief is written. The five sample leads
19:50are pulled with phone and email. the
19:52other 45 leads and the five static ads
19:54and two videos have not been started.
19:57Are these five the right kind of leads
19:58so I can pull the other 45 and make the
20:00ads and videos? I'm going to say yes,
20:02continue.
20:04And actually, what I'm going to say is
20:05yes, continue, but make it 20 total
20:09leads because again, for the sake of
20:11this video, I think we understand the
20:12power of what Fable can do here. And
20:15it's not super necessary to pull all 50
20:17leads right now. Let's just make it 20
20:19so that we can get on with the episode.
20:21And I I'm really interested to see the
20:23ads. Like that's what I want to see is
20:25are these ads going to be high quality?
20:27Are these ads going to look like crap?
20:29We'll see. But I'm going to say
20:30continue. Make it 20 total leads. And I
20:32will be back when it has the uh
20:35remaining 15 leads that it's going to
20:37pull the five static ads and the two UGC
20:39videos. All right, it's done. I haven't
20:42looked at the ads or the UGC videos yet,
20:44so you're going to be seeing it live
20:45with me. Um, this is so this looks like
20:48the first UGC video. This is the second
20:51one. And then these are the actual ads
20:54themselves. So
20:56these actually look pretty good
20:58considering we have no references to go
21:00off of, just the style guide.
21:03I kind of like these. So let's let's
21:05listen to the UGC ads first. We'll just
21:07listen to Well, let's listen to one.
21:09>> If you own a restaurant, do this math
21:11with me. Say you do $8,000 a month
21:13through the delivery apps. At 25%
21:15commission, that's $2,000 gone every
21:18month for food you bought, you prepped,
21:20and you cooked. Your regulars will order
21:22from you directly if you give them away,
21:24too. Mesa Direct builds your own
21:26ordering site in 7 days. We load your
21:28menu, connect your payments, and you pay
21:30one flat monthly fee, zero commission.
21:32Keep the apps for new customers. Move
21:34your regulars to your own site. The link
21:36is in the bio.
21:37>> So, not bad. It sounds a little AI
21:39generated in the sense that there's not
21:41a ton of emotion, but for a first pass,
21:44like that's pretty good. Obviously, we
21:46could go in and and do a few things to
21:47touch it up and make it even better. But
21:49for a first pass, one shot attempt with
21:52Fable 5.2 using that pre-installed skill
21:56from Graph, like literally all I did was
21:58install that skill. I didn't do anything
21:59else other than give it that skill and
22:02tell it to go to town. I think that's a
22:03pretty good first take. And with the
22:05second one here, again, I won't sit here
22:06and listen to the second one, but I
22:08think probably similar to the first and
22:11pretty good first attempt. Now, the
22:13actual ads themselves,
22:15uh, you cooked it, they took up to 30%.
22:17Take orders on your own site and pay 0%
22:19commission. See what you keep. That's
22:21the call to action. Stop paying
22:23commission on your regulars. Uh, the app
22:26knows your regulars to you. Every bag
22:28can bring the next order.
22:31Your own ordering site live in seven
22:32days. So, I I like think these ads are
22:35really solid all things considered,
22:36especially because again, we didn't give
22:38it a website URL that it went and built
22:40these off of. It generated the style
22:43guide and then it created the ads based
22:44on the style guide. Again, using that
22:47free skill from Graph and I'll put those
22:49uh a link to those Graph skills in the
22:50description. Again, I have no
22:52affiliation with them. I just think that
22:53this has been really helpful in helping
22:55us do this test. So, in my mind, again,
22:59the UGC ad probably like a six or seven
23:01out of 10. definitely needs a few
23:03iterations to clean that up and get it
23:04to where it's like truly top-notch and
23:08I'm willing to ship it. But the static
23:09ads I think are solid and I would have
23:11no issue shipping these static ads as is
23:14and at the very least testing them. And
23:16I'm sure we would start to get uh a few
23:18inquiries or a few leads within
23:20definitely within the first week just
23:22from these static ads. And again, it
23:24literally cost us 20 cents give or take
23:27a couple of pennies to generate these
23:29five static ads. We could generate 100,
23:32200, 500 variations of these ads and
23:35test them all at once. And 100% we would
23:37have multiple leads within the first
23:40week from doing just that. Now, it uh it
23:43stored the static ads on my local
23:46machine, which I love. It stored the UGC
23:48videos as well uh both in my graft
23:51account and on my local machine.
23:53Remember, it found us 15 additional
23:55leads. And so all in all, this cost us
23:5930 cents for the leads via Tre. It cost
24:02us about $9 via Graph for to use 900
24:07graph credits for the two UGC videos
24:09plus the five images because the five
24:11images cost us 3.5 each. Um, so
24:15everything that we just did cost us
24:18$930.
24:20Now it said, you know, make sure the
24:22name Mesa Direct isn't trademarked. We
24:24haven't built a landing page yet. And
24:26both videos end by pointing to a link.
24:28So what Fable wants to do for us next,
24:31it says, "Hey, should I build the
24:32landing page that the ads and the videos
24:34point to next?" So we can just press a
24:37button and it's going to go and oneshot
24:39a landing page that I'm sure is going to
24:41look pretty dang good because Claude has
24:43always been really good at front-end
24:44design. It says, you know, build landing
24:46page or just write the outreach uh
24:48scripts. So I'm actually going to call
24:50it there because this is everything that
24:52we promised we would do during the
24:53video. So, we're 24 minutes and 55
24:55seconds in and we have uh literally
24:59built a full business from the ground
25:01up. We came up with an idea. We created
25:04the offer. We found the ICP. We found 20
25:06leads. We generated five ads. We
25:09generated two UGC ads. And we could next
25:12build a landing page, write outreach
25:13scripts, do all of this a 100 times over
25:16or do this for 50 different offers. like
25:18that. It's truly mind-boggling how easy
25:21it is to get up and running these days.
25:23Now, obviously, the hard work is in
25:25doing the outreach, making the calls,
25:26sending the emails, and that'll be the
25:28topic of a separate video, and I'm happy
25:30to go deep there. So, if you guys
25:31enjoyed this, be sure to leave a comment
25:33and subscribe. We built a full business
25:35in a box in under 26 minutes. Hope you
25:37guys enjoyed and I'll be back

About this transcript

This page contains the full transcript of make f*ck you money with claude fable 5.2 + free github repos by Corey Ganim, generated from the public captions YouTube serves with the video. The transcript has 5,118 words across 695 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.