I've been building a lead generator for a client and Apify is the engine underneath it. It's a store of pre-built scrapers, thousands of them, that you rent by the run instead of writing and maintaining your own.
Two things about it made it worth a post.
The first is the price. Pulling five welding shops in Reno with contact details attached cost about two cents. Four of the five came back with a company email on them. Everyone assumes scraping at any real volume is a serious line item, so it's worth seeing the actual number.
The second is the bit I'd actually steal. I connected it as an MCP server rather than hardcoding one scraper into a script, which means Claude searches the store and picks the scraper at runtime. When one gets deprecated or starts failing, the agent goes and finds the replacement itself instead of the whole pipeline falling over at 2am.
Here's the part most people skip though. Scraping the list is the easy half. I added a second actor that verifies every email before anything gets sent, at a tenth of a cent each. It immediately caught a filler address sitting on a godaddy domain plus a generic service inbox that would have bounced. It also collapsed four locations of the same chain into one company, so I wasn't mailing a national brand four times while counting it as four leads.
Getting the data is now cheap enough that it stopped being the hard problem. Working out which rows are junk is the hard problem now. That's the step almost everybody leaves out because it isn't the fun bit.
Think about the last list you scraped or bought. Did you verify any of it before you sent?