You can give Claude Code live access to the web and have it gather leads for you, using a free, open-source tool called Firecrawl. You describe the kind of business you're after, and Claude searches the web and reads the pages it finds, then hands back a clean list of companies with their public contact details. Here's the full setup, the prompts, and how to run it without getting yourself blocked or breaking any rules.
Firecrawl is the official web-scraping plugin for Claude Code (github.com/firecrawl/firecrawl-claude-plugin). It turns any page into clean data Claude can use, handles JavaScript-heavy sites, and is built compliance-first. The free tier gives you 500 credits to start, which is plenty to test before you pay for anything.
The quickest route is the plugin. In Claude Code, run:
/plugin
Search for firecrawl, select it, and install. It needs the Firecrawl CLI installed globally and a free API key from firecrawl.dev/app.
Prefer the MCP route? One command does it:
claude mcp add firecrawl -e FIRECRAWL_API_KEY=your-api-key -- npx -y firecrawl-mcp
Now just describe your target in plain English. A prompt like this:
Use Firecrawl to find me 25 pool-cleaning businesses in Austin, TX.
For each one, pull the business name, website, publicly listed email and phone, and two specific details from their site I could mention in outreach.
Give it to me as a clean table.
A couple of minutes later you've got a usable list, pulled from the businesses' own public pages.
Firecrawl gives Claude more than basic scraping, so you can: