n8n Google Maps scraper workflows

The templates

Set it up

1. 1. Import the workflow. In n8n, open Workflows → Create → ⋯ → Import from File and pick the JSON (or paste its contents onto the canvas). The yellow sticky notes repeat these steps.
2. 2. Get an API key. Sign up (100 free credits, no card), then copy a key from Dashboard → Settings. It starts with hub_live_ and is shown once.
3. 3. Add the credential. Create one Bearer Auth credential with the key as the Bearer Token, and select it in the three HTTP Request nodes: Start scrape, Get scrape status and Fetch all rows. Bearer Auth arrived in n8n 1.88; on older versions use a Header Auth credential instead and set each of those nodes to Authentication → Generic → Header Auth:

   Name:  Authorization
   Value: Bearer hub_live_your_key_here

4. 4. Connect the destination. Google Sheets via OAuth2, pointed at a spreadsheet with a Leads tab (or Places for the weekly monitor; the header row is written on the first run), or HubSpot via a private-app token.
5. 5. Edit the search and run. Open Start scrape and change the JSON body, then execute the workflow (or activate the weekly monitor).
   Copy

   {
     "categories": ["dentist"],
     "locations": ["Austin, TX, United States"],
     "maxResults": 50,
     "enrichWebsite": true,
     "enrichLinks": true
   }

   Add "country": "gb" to pin a Google country edition; otherwise it's read from the locations. Every field is described in the API docs.

Node by node

All three share the same scrape-and-poll block after the trigger:

1. 1. Start scrape — HTTP Request, POST /api/v1/scrapes with your categories, locations and maxResults. It doesn't retry, so a timeout can't start the same scrape twice.
2. 2. Wait 30 seconds — Wait node.
3. 3. Get scrape status — HTTP Request, GET /api/v1/scrapes/{id}. Retries 3 times, 5 seconds apart, on transient errors.
4. 4. Scrape finished? — IF status is done, go on to fetch the rows; otherwise check for failure.
5. 5. Scrape failed? — IF status is error, Stop and Error with the API's message (the reserved credits are refunded). Otherwise loop back to Wait 30 seconds.
6. 6. Fetch all rows — HTTP Request, GET /api/v1/scrapes/{id}/results with n8n's built-in pagination: 1,000 rows a page until hasMore is false, up to 100 pages (100,000 rows).
7. 7. One item per place — Split Out on rows: one n8n item per business.

What it costs in credits

• • 1 credit per business delivered. Each run reserves categories × locations × maxResults (capped at your balance) and refunds what the scrape didn't use. Emails and social profiles cost nothing extra. On Starter that's $0.38 per 1,000 businesses; see pricing.
• • Filters don't save credits. Has email? runs after delivery, so businesses without an email are still charged.
• • The weekly monitor spends every week. It re-scrapes the whole search each Monday: 1 category × 1 city at maxResults 200 is at most 200 credits a week (about $0.08 at the Starter rate). Keep maxResults modest.

Limits

• • maxResults is 1–500 per category × location pair, and one scrape covers up to 2,000 pairs.
• • 12 new scrapes per 5 minutes per account, so don't fan one workflow out into dozens of parallel runs.
• • Fetch all rows stops at 100 pages, 100,000 businesses per run.
• • The Google Sheets node reads columns A–Z, so the templates write 20–21 chosen columns. The full 38-column file is one call away: GET /api/v1/scrapes/{id}/export?format=csv.

Questions

Prefer a spreadsheet without n8n? The Google Sheets script does the scrape-and-dedupe inside Sheets. All integrations: /integrations.