← back
scrapewithclaude.com
Claude can give you superpowers, and most people have no idea what it's actually capable of. Scraping is one of them. At its core it's just telling Claude to go out on the internet and find you something. A list of companies. The software they run. The people who work there.
You give it a few tools and a little bit of instruction, and what comes back will blow you away.
So I wrote the whole system down as a playbook and built a site that teaches it. All of it free.
scrapewithclaude.com →
golf-scrape · terminal
$ claude
scraping golf-courses.csv · overnight run
→ scraping pebblebeach.com ... ✓ contact found
→ scraping erinhills.com ... ✓ contact found
→ scraping chambersbaygolf.com ... ✓ contact found
verified contacts: 312 / 400 avg cost: $0.029
sample from a real run, contacts masked
It's for the people whose job is finding customers: marketers, sales, rev ops. This space is brand new and moving fast, and almost nobody is sharing the really good stuff. The people who figure it out keep it quiet. I think it's more fun to give it away.
The centerpiece is the playbook. One story, a software company selling to golf courses, carried through the three reasons anyone scrapes: build the list, find the signal, feed the outbound. Then seven techniques for getting data off any website, three copy-paste playbooks, and a five-minute setup. Real screenshots, real run logs, real costs.
Around it, the site does the quiet work. Every technique and every tool gets its own page, each one aimed at a keyword somebody actually searches. Those pages are the way in. The playbook is where they land, and the email list is what compounds.
What's in the playbook
Three reasons to scrape, one story
A booking-software company works 16,000 golf courses from flat list to ranked pipeline to outbound.
The signal chapter
What a prospect's own website tells you that changes the pitch. The argument against fake personalization.
Three copy-paste playbooks
Find the companies, find the contacts at three cents a verified email, run the outbound.
A five-minute setup
One paste into Claude Code and the scraping machine is built. Claude installs what's missing.
The plays
What software is this company running?
Read a site's page source and tell which tools it runs. Shown on dentists, works anywhere.
Map a tool's entire customer base
Backlink and subdomain footprints hand you a SaaS product's full customer list.
Find the hidden subdomains
Certificate logs and wordlists surface the endpoints that aren't in the nav.
Catch the JSON the site already loads
Most dynamic sites fetch clean JSON, then draw it. Grab the JSON, skip the page.
Read the raw HTML, let the LLM parse it
One request, one prompt, fields by meaning instead of brittle selectors.
Scrape the sites that fight back
A headless browser renders the JavaScript-only pages that a plain request can't see.
Unlock the ones that still block you
A residential proxy gets past hard blocks. The last and most expensive tier.
How it's built
Stack
Pages
Static HTML, Tailwind tokens, Netlify
Gate
Netlify form, a wall on the book, a transition that makes it feel intentional
Fetch tiers
Plain request → headless browser → stealth browser → residential proxy
SEO
One page per winnable keyword, plays and tools both
The idea behind it
Free
Everything is free to read. The book costs an email.
Tactical
Real tools named, real prompts shown.
A funnel
Pages rank, the playbook converts, the list compounds.
A drumbeat
Built to add plays and tool pages over time.
Note
The sample data on every play is anonymized. The techniques are the point, not anyone's data. The whole thing is free to read and free to copy.