finding-local-businesses

Scrape local businesses from Google Maps with addresses, phones, websites, emails, and decision-maker contacts. Use when the user wants local leads, businesses in a city, contractors/restaurants/salon

SKILL.md

GitHub

Finding Local Businesses (System 2)

Goal

Turn "barbers in Tuscaloosa" into a spreadsheet of businesses with contact details — and, if asked, the named decision-makers inside them.

When to use

  • "Find me every roofer in Mobile Alabama"
  • "I need local restaurants with emails"
  • "Scrape Google Maps for HVAC companies"
  • Anything the user calls System 2, local leads, or Google Maps

The one thing to know

The actor got much better since the course was recorded. The video needed six steps to get emails: scrape Maps, loop over results, fetch each website, AI-extract the contacts, export a CSV, upload it to AnyMailFinder. That's now three flags on one command:

Flag What it adds
--contacts Emails and social profiles pulled from each business website
--leads N Up to N decision-makers per business: name, job title, email, phone, LinkedIn
--verify-emails Checks each of those emails is real before you send

Each one bills separately. Start without them.

Process

Step 1 — cheap test, no add-ons

python -X utf8 scripts/02_local_businesses/scrape_google_maps.py \
    --search "barbers in Tuscaloosa AL" --limit 5 --out-csv output/test.csv

About 2 cents. Read the CSV — right businesses? right city?

Step 2 — the real run

python -X utf8 scripts/02_local_businesses/scrape_google_maps.py \
    --search "roofing contractors in Mobile AL" --limit 100 \
    --contacts --out-csv output/roofers.csv

Step 3 — decision-makers, when the user needs a name to write to

python -X utf8 scripts/02_local_businesses/scrape_google_maps.py \
    --search "law firms in Birmingham AL" --limit 50 \
    --contacts --leads 2 --departments "C-Suite" --verify-emails \
    --out-csv output/lawyers.csv

Full pipeline with AI summaries

When the user wants a 50-word summary and 5 facts per business (the video's flow):

python -X utf8 scripts/02_local_businesses/gmaps_lead_pipeline.py \
    --search "dentists in Miami FL" --limit 25 --out-csv output/dentists.csv

This adds summary and fact_1..fact_5 columns. Google Sheets is opt-in via --sheet-url; without it you get CSV and no OAuth setup.

Useful filters

  • --min-stars 4 — skip businesses with poor ratings
  • --skip-closed — drop permanently closed places
  • --actor crawler — richer data (reviews, hours) for more money; default extractor is what the course used and rates highest

Edge cases

  • Search term matters more than anything. "barbers in Tuscaloosa AL" beats "barbers" + --location. Put the city in the search string.
  • One location per run. Two cities in one search returns mush. Loop instead.
  • ~15% of business websites block scrapers — --contacts will come back empty for those. Normal, not a bug.
  • Fewer results than --limit — Google Maps genuinely has that many. Broaden the term or widen the area.
  • --verify-emails without --leads does nothing; verification only applies to the decision-maker records.

Cost

~$0.004–0.02 per place, plus each add-on. A 100-business run with contacts is usually under $1. See docs/COSTS.md.