Feb 16, 2026 · 55m · startup-ideas
Claude Code built me a $273/Day online directory
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Host Greg Isenberg and guest Frey Chu break down how to build profitable online directories using Claude Code and Crawl4AI to automate data curation and capture programmatic search traffic. Through real-world case studies and a live build of a luxury restroom rental directory, they reveal actionable scraping pipelines, monetization strategies, and techniques for thriving in the age of AI search.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Greg holds 18.4% of the talking time here. How this is scored →
Greg as informed peer 3.8 Guest teaching 6.2 Guest disagreement 0.2 Greg pushing back 1.1
speaking balance: gold is Greg, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 44:25 Frey's pushback on get-rich-quick expectations
Frey directly counters the hype of instant monetization, telling listeners needing money in under six months to flip garage sale items instead of building directories.
Hardest push from Greg ▶ 33:47 Greg's legal warning on scraping imagesGreg directly interrupts Frey's image workflow to clarify the legal risks and terms-of-service violations associated with scraping images without explicit licenses.
Biggest teaching moment ▶ 26:25 Automating 1,000 hours of manual verificationFrey details how running Crawl4AI with async web crawling via Claude Code replaces thousands of hours of manual verification for under $50 in API credits.
Greg holds their own ▶ 52:29 Greg's synthesis on AI search citation dynamicsGreg demonstrates sharp platform expertise, articulating how specific LLM prompting will bypass generic horizontal aggregators and cite niche single-purpose directories.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Greg as informed peer | Guest teaching | Guest disagreement | Greg pushing back | Why |
|---|---|---|---|---|---|---|
| Guessing Game: Analyzing Profitable Directory Models | 4 | 5 | 0 | 0 | Greg participates in a guessing game regarding parting.com's traffic and monetization model. Frey reveals that rather than just lead gen, Parting generates the bulk of its revenue via a vertical SaaS cremation arrangement tool and agency backend. | |
| Case Study 2: A Place for Mom and High-Ticket Lead Gen | 5 | 4 | 0 | 0 | Greg demonstrates industry knowledge regarding senior living real estate investment trusts and their marketing spend. Frey shares specific metrics on A Place for Mom, noting over 800k monthly organic visits and $50M+ revenue estimates. | |
| Case Study 3: GasBuddy, Gamification, and Price Transparency Moats | 5 | 6 | 0 | 0 | Greg cites his own agency's grocery data project Stretch AI to highlight data moats. Frey walks through GasBuddy's 20-year gamification history with leaderboards and debit cards, teaching the core thesis of price transparency as a directory moat. | |
| The Porta-Potty Experiment and High-Ticket Inbound Demand | 2 | 7 | 0 | 0 | Frey presents his initial porta-potty directory experiment, demonstrating how an ugly MVP with placeholder text still generated a $20k+ state fair inbound order, convincing Greg of the niche demand. | |
| Step 1 & 2: Scraping Raw Data and Initial Cleaning | 1 | 7 | 0 | 0 | Frey delivers a technical tutorial on scraping 71k listings using Outscraper and prompting Claude Code to strip out junk listings down to 20k rows. | |
| Step 3: Deep Verification Using Crawl4AI and Claude Code | 2 | 8 | 0 | 0 | Frey demonstrates integrating Crawl4AI's async web crawler with Claude Code to automate website verification in 3 hours instead of 1,000 manual hours. Greg highlights that Frey achieved this despite being non-technical. | |
| Step 4 & 5: Inventory Enrichment and Claude Vision Image Verification | 6 | 7 | 0 | 4 | Frey explains data enrichment prompts and using Claude Vision to filter scraped images. Greg intervenes with legal pushback, reminding viewers about copyright risks and terms of service when scraping images. | |
| Step 6 & 7: Feature Filtering and Service Area Mapping | 3 | 6 | 0 | 0 | Greg playfully notes the loss of character from the older MVP's logo. Frey demonstrates feature extraction (running water, stalls) and mapping service radiuses into Supabase tables. | |
| Identifying High-Value Niche Directory Ideas and Public Datasets | 2 | 7 | 0 | 0 | Frey educates listeners on finding high-ticket sub-niches (e.g., dementia senior care, ADA contractors) and mining public datasets on data.gov for tap water quality directories. | |
| Evaluating Directories as Distribution-First Business Models | 5 | 6 | 1 | 3 | Greg asks Frey to address naysayers who doubt directory viability. Frey gives a pragmatic reality check that directories are not for people needing cash within 6 months, framing them instead as distribution-first sandboxes for AI coding and SEO. | |
| The Impact of AI Search and LLMs on Online Directories | 7 | 5 | 1 | 5 | Greg presses Frey on whether LLMs and Perplexity will cannibalize directory traffic. Frey defends the decision-making stage and local SEO; Greg builds on this with his thesis that AI search specifically rewards hyper-niche directories over horizontal ones. |
Statements from this episode (24)
Chu: Acquiring valuable data is the hardest part of building AI directories
“Yeah, by the end, I really want people to have a solid understanding of how to use cloud code to build AI coded directories with an emphasis on what I think is the hardest part, which is getting the valuable data.”
Frey Chu: Parting.com generates over $1M annually and raised $1.5M
“This is reported due over a million dollars a year in revenue. There's some quotes that go up to five million dollars, but they even raised money. They raised 1.5 million dollars in 2021.”
Frey Chu: Pairing directories with vertical SaaS or agencies is highly effective
“I still believe that building a directory and creating a micro SAS or vertical SAS is a super solid way to go about it. And with cloud code, it's never been easier. But the agency directory combo is a potent one and I see it all the time. Some of the biggest d…”
Chu: APlaceForMom.com likely generates around $50 million in annual revenue
“My gut tells me like fifty million dollars a year is probably right where I feel like it's more, more accurate.”
Chu: APlaceForMom.com takes first month's rent or fixed fees for referrals
“And they basically have all of these kind of sneaky ways to collect information and then pass them off to these senior living homes where they collect a portion of the first month's rent or like a fixed fee for every referral that they give.”
Chu: Data and SEO backlinks are the primary moats for online directories
“I would say for directories, the moat is definitely data and your SEO, like, if you have really strong backlinks.”
Chu: Luxury restroom trailer rentals cost $1,000 to $2,000 per day
“People are spending a thousand dollars, 2000 dollars a day renting these out for their weddings.”
Chu: Unfinished porta-potty directory generated a $20,000+ New Mexico State Fair lead
“And then this is the big one. I got a lead from the New Mexico State Fair. This is the order that they wanted. And this is like 20,000 dollars plus of porta potty slash luxury restroom trailers that they wanted.”
Chu: Built luxury restroom directory in 4 days for under $250
“I mean, I built this in four days. Like I mentioned, I probably saved over 2000 hours in what would have been manual data cleaning, manual data enrichment, also record low costs. I mean, I would have to hire a developer to like write custom Python scripts befo…”
Chu: Outscraper is the cheapest option for scraping Google Maps data
“I just went straight to outscraper. I don't think. There's nothing new here. I went over this in the last pod, but Altscraper is still the cheapest option for me. I've used Appify and other kind of alternatives, but just keep it simple. I think Altscraper is a…”
Chu: Initial Claude Code prompt filtered 71,000 listings down to 20,000
“And so this simple prompt got me from 71 K down to 20,000 listings.”
Chu: Crawl4AI is a free open-source crawler for LLMs
“This is an open source LLM friendly web crawler and scraper totally free, which is insane.”
Chu: Claude Code filtered 20,000 sites to 725 listings in three hours
“And then it just ran in the background for three hours. And this is really what would have taken me like a thousand hours just a couple of years back. So the result was. 20,000 websites down to 725 luxury restroom trailers.”
Chu: Single-pass LLM data enrichment fails compared to sequential data extraction
“And one of the first mistakes that I did when I started enriching my data with Crawford AI was I would just give it this massive laundry list of things to get. And just gave it like a massive CSV. And I was like, get the trailer inventory, get the images, get …”
Chu: Claude Vision verified 700 image listings for $30 in one hour
“This cost about 30 dollars. All I had to do was go to Claude and connect my API key. Super simple. And this is the prompt that I used. I let this run overnight, but yeah, this took probably about an hour or so for 700 listings.”
Chu: Online directories do not need listing images to rank in search
“And I think you don't even need images for a directory to rank and get traffic.”
Chu: Targeting dealbreaker long-tail features is the best directory SEO strategy
“And I think this kind of idea works well because for the people looking for this, like if they don't assist with dementia, that's a deal breaker. So what are like the deal breaker features that people make decisions off of? And how can we create like very spec…”
Chu: Zero-backlink tap water directory reached 40,000 monthly visitors in months
“This guy named Andy in our community, but a tap water quality directory, no backlinks started it back in, I want to say November and he's getting over 40,000 monthly visitors got accepted into the media vine selling like water filters through Amazon affiliate.”
Chu: Avoid building directories if you need revenue within six months
“If your timeline is To make money in less than six months, I would not build a directory.”
Chu: LLMs dominate discovery, but directories capture high-stakes decision-making
“What I think ChatGPT perplexity and all these other LLMs do really well is they help you discover things or get more information on the thing. But by the time someone's on a directory, I think they're in the decision-making phase and with any more complex deci…”
Chu: Avoid product-based niches when building directories due to heavy competition
“This is the reason why when I build directories, I don't choose product-based niches because it's way harder. You're just competing with so much other content and like socials and all this other stuff.”