> ## Content Index > Fetch the complete content index at: https://www.theopportunitydesk.blog/llms.txt > Use this file to discover other available public pages before exploring further. # Greg Isenberg & Frey Chu: Claude Code Built me a $273/Day online directory - URL: https://www.theopportunitydesk.blog/greg-isenberg-frey-chu-claude-code-built-me-a-273-day-online-directory/ - Published: 2026-06-23T23:06:31.000Z - Updated: 2026-09-03T21:02:14.000Z - Description: Frey Chu says Claude Code and a web scraper let him build a niche directory in 4 days for under $250, with traffic arriving "on autopilot." The pitch skips the six-to-twelve-month SEO grind, and he never discloses whether his own directories have turned inbound leads into revenue. - Author: The Opportunity Desk - Tags: channel: greg-isenberg, category: niche-websites, guest: frey-chu --- Source: "Claude Code built me a $273/Day online directory" — Greg Isenberg, published February 2026\. [Watch on YouTube →](https://www.youtube.com/watch?v=I%5Fwbc5ND79o&ref=theopportunitydesk.blog) --- **THE OPPORTUNITY** Frey Chu runs a consulting and education business teaching people to build niche directories - websites that compile listings of local service providers in a single category. He demonstrates a workflow using Claude AI and an open-source web scraper called Crawl4AI to take messy data and turn it into structured directory pages. The pitch: build a directory in a few days for under $250, rank in search results through topical authority, and monetize through lead generation or vertical software products once traffic arrives. --- **HOW IT'S EXECUTED** 1. Scrape raw business data from Google Maps using a paid tool like Outscraper (covers entire states or countries, generates tens of thousands of rows). 2. Use Claude Code to remove obvious junk such as closed businesses, listings with missing contact details, and irrelevant entries like big-box retailers. 3. Install Crawl4AI locally and connect it to Claude Code to visit each remaining business website and verify that it matches the niche (e.g., luxury restroom trailers, not standard porta-potties). 4. Run additional Crawl4AI passes to extract specific details from each verified business's website - product inventory, amenities, service areas, high-quality images. 5. Use Claude Vision API to evaluate scraped images and select the best ones (costs additional API credits). 6. Feed the cleaned, enriched data into a database (Supabase in Frey's case) and use Claude Code to generate the directory frontend from that database. --- **WHAT'S CREDIBLE** Frey shows receipts for an earlier directory he built - PortaPottyMatch.com - that generated real inbound leads despite placeholder text and identical stock images on every listing. One lead was from the New Mexico State Fair requesting over $20,000 worth of rental equipment. That validates the underlying demand: businesses and event planners do need comparison tools in this category, and a directory with poor execution still pulled qualified leads. The three case studies at the top - Parting.com, A Place For Mom, GasBuddy - are real directories with verifiable traffic and revenue models. Outscraper and Crawl4AI are real tools with documented pricing and open-source repos. --- **WHAT'S OMITTED OR OVERSTATED** He says the directory was built in four days. He does not specify whether that was four calendar days or four days of concentrated work spread over a longer period. He does not explain how much iteration and debugging happened inside those four days. The "$250 total cost" includes a $100 Claude subscription, $100 for data, and $50 for API credits. It does not include the cost of a domain, hosting, or any backlink-building expenses required to actually rank the site in search results. He says the site will get "traffic on autopilot" once it ranks, but he never explains how long it takes to rank or what work is required to get there. SEO timelines for new domains without existing authority can easily stretch beyond six months. He acknowledges this obliquely near the end when he says "if your timeline is to make money in less than six months, I would not build a directory," but that warning comes after an hour of framing the process as fast and passive. He mentions he "probably saved over 2,000 hours" by using this workflow. That number is unsourced and implies a comparison to manual work he never actually did at that scale. He ran his older directory manually on a much smaller dataset, so the time-savings claim is speculative. He scrapes images from business websites and admits "it's a gray area." He says he plans to reach out and ask permission retroactively by having businesses claim their listings. This is not how copyright works. You need permission before use, not after. Posting scraped images without licensing rights exposes the site owner to DMCA takedowns and potential legal claims. Stock images or user-uploaded photos would avoid this entirely. He never discloses how much he has earned from any directory he has built. The case studies he shows are other people's businesses, not his own. The lead screenshots from PortaPottyMatch.com prove inbound interest but do not prove he closed any deals or collected any revenue. He says directories monetize through lead generation, ads, or vertical SaaS. He does not explain what a lead is worth in the luxury restroom trailer niche, what close rate to expect, or what profit margin remains after paying for traffic acquisition and lead nurturing. He says GasBuddy monetizes through ads and a premium debit card but does not explain how a new directory with no brand recognition or card partnership would replicate that model. He says "lead generation will always be around" as a business model. He does not address how AI search engines scraping and summarizing directory data could devalue the directory itself as a destination. He briefly counters this by saying users in a decision-making phase will still visit directories to compare options, but he provides no data to support that assumption in an AI-first search landscape. He says local SEO has not changed much and remains low-hanging fruit. He does not mention that Google's local pack and map results often capture the majority of clicks, leaving organic directory listings below the fold with minimal traffic. He says building a directory teaches you Claude Code, SEO, and lead generation. He mentions running a free directory community with over 3,200 members, but it's unclear whether he also sells paid courses, coaching, or premium tiers beyond that community - a detail that would matter for assessing whether his incentive is to make the model look accessible regardless of individual outcomes. He says you can build "nationwide directories" and implies broad coverage is an advantage. He does not explain that ranking for competitive terms in major metros requires significant domain authority, backlinks, and time -advantages a brand-new directory does not have. --- **BOTTOM LINE** This is a valid workflow for structuring and enriching messy data using AI tools, and it works if you already know how to rank a site and monetize traffic. It is not a passive income shortcut. You need SEO skills, patience to wait six to twelve months for rankings, and either sales skills to close leads or existing relationships in the niche to build a vertical SaaS. Frey's older directory pulled leads, but he does not disclose whether those leads converted to revenue or profit. The scraped-image approach creates legal risk. If you lack SEO experience or a specific niche with weak competition and strong buyer intent, this will consume time and capital without returning income. --- **OPPORTUNITY DESK RATING** **Feasibility: 3/5** — The underlying business model works (comparison directories do generate leads in service categories), but the "four days for $250" framing hides the SEO timeline and backlink costs required to actually get traffic, and the revenue path is unproven in Frey's own case. **Transparency: 2/5** — He shows real tools and a real workflow but omits his own earnings, underplays the legal risk of image scraping, and frames a six-to-twelve-month SEO grind as "autopilot traffic" without explaining the ranking work required upfront. --- ---