Most lead generation tools charge per record, per export, or per seat. Apollo charges by contact unlock. Apify charges by compute unit. Clay charges per enriched row. ScrapeGraphAI charges nothing. It is a free MIT-licensed Python library with 26,000+ GitHub stars that lets you describe in plain English what to pull from any page and returns structured data. Pair it with Claude Code as the orchestrator and you have a lead pipeline that runs for the cost of electricity.
Install
pip install scrapegraphai
playwright installPlaywright handles JavaScript-rendered pages that basic scrapers miss. Run both commands before your first scrape.
Basic Usage — Describe It in Plain English
Open Claude Code and describe what you want to extract. Claude Code writes the ScrapeGraphAI script, runs it, and returns your data. No Python required for the initial run.
Scrape https://example-directory.com/listings and extract: company name, contact email, phone number, city, and business category. Return the results as a CSV saved to ~/Desktop/leads.csv. Use ScrapeGraphAI for the extraction.Swap in any URL — business directories, LinkedIn search results, conference attendee pages, job boards, real estate listings, local association member directories. ScrapeGraphAI's AI extraction layer handles page structure differences so you do not write custom selectors for every site.
Layer 2: Verify and Enrich Before You Send
Raw scraped email lists bounce at roughly 10–15%. Running verification before any campaign pushes deliverability above 90% and protects your sender reputation.
- Email verification — upload your CSV to MillionVerifier or ZeroBounce. Both have free tiers. Remove any record with a deliverability score below 70.
- Deduplication — ask Claude Code to deduplicate by email domain and company name. One prompt, no manual sorting.
- Enrichment — pass company names through Hunter.io's domain search or Clearbit's free tier to fill in missing titles and LinkedIn URLs.
- Segmentation — ask Claude Code to split the verified list into segments by industry, city, or company size based on what was scraped.
The gap between a raw scraped list and an outreach-ready list is verification. Skip it and high bounce rates degrade your sender domain — your next campaign pays that cost even if its list is clean.
What It Replaces
- Apify — ScrapeGraphAI handles the same extraction with zero compute billing and a smaller setup surface
- Apollo.io for list building — targeted scraping from directory pages often pulls equivalent data for free
- Lead-gen VAs — a Claude Code workflow running ScrapeGraphAI + MillionVerifier runs overnight unattended
- Clay for enrichment — for moderate volumes, Hunter.io's free tier plus Claude Code enrichment covers what Clay charges per enriched row
Use Cases Beyond Lead Lists
- Competitor pricing — scrape competitor product pages on a schedule and log every price change to a spreadsheet
- Real estate — pull active listings from local MLS aggregators or Zillow search results by ZIP code
- Job monitoring — scrape Indeed or LinkedIn Jobs for roles matching a skill set and pipe them into a Notion tracker
- Press and mention monitoring — scrape Google News results for a brand name or keyword and send yourself a daily digest
ScrapeGraphAI lives at github.com/ScrapeGraphAI/Scrapegraph-ai. The MIT license means you can run it commercially, embed it in client deliverables, or build a service on top of it. When you are paying per record for lead data, the first question worth asking is whether ScrapeGraphAI plus 30 minutes of Claude Code setup gets you the same list for free.