Full automation pipeline: scrape Google Maps businesses → filter by website presence → export to Google Sheets. Built with n8n and OutScraper.
Finding local businesses that don't have a website is entirely manual by default. You search on Google Maps, scroll through results, check each listing one by one, and note down the name, category, and contact info — business by business.
That process is slow, repetitive, and doesn't scale.
The goal: type a search like "Italian restaurants Paris 11th" and instantly get a structured list of businesses with no real website, ready for outreach.
A lightweight web interface built in HTML / CSS / JavaScript — 3 files, nothing more.
The user:
- Types a Google Maps search query (e.g.
Plumbers Lyon,Italian restaurants Paris 11) - Sets the number of results to extract (1–500)
- Clicks "Start extraction"
- Gets a direct link to a Google Sheet with all the data
An animated progress bar runs during extraction, and the link appears automatically once the job is done.
The workflow is triggered by a webhook called by the interface. Two parallel branches then run simultaneously:
- Branch 1 — Creates the destination Google Sheet with the correct headers, ready to receive data
- Branch 2 — Starts the scraping job via the OutScraper API
Running both in parallel saves time: the spreadsheet is already set up while the scraping is still running.
The OutScraper API doesn't return data immediately. It works with a task ID: you have to poll the API in a loop until the job is complete.
A naive fixed wait of 30 seconds creates two problems:
- Too long → bad user experience
- Too short → the workflow fails
The solution: the entire polling logic is extracted into a dedicated sub-workflow. The main workflow calls it via an Execute Workflow node and stays blocked until a result comes back.
The sub-workflow handles all the complexity:
- Polls the API every second
- Increments an attempt counter
- Checks two exit conditions:
- ✅ Status is
success→ returns data to the main workflow - ⏱️ Exceeds 60 attempts → returns a clean
timeoutstatus instead of an unhandled error
- ✅ Status is
Key benefit: this polling sub-workflow is fully reusable across any project that uses an async API — OutScraper, Apify, or others.
Once the data is retrieved, a filter is applied on the website field:
Businesses whose only web presence is a Facebook, Instagram, or Tripadvisor link are flagged as priority leads — they have no real website.
The filtered data is then:
- Injected into the Google Sheet created earlier
- The sheet is automatically shared with public read access
- The URL is sent back to the interface so the user can access the list immediately
| Before | After |
|---|---|
| ~10h/week of manual research | A few seconds per extraction |
| One city at a time | Multiple cities in parallel |
| Data written down by hand | Structured Google Sheet, ready to use |
| Non-reproducible process | Automated, reusable workflow |
What used to take half a day now runs in seconds.
| Layer | Technology |
|---|---|
| Frontend | HTML, CSS, JavaScript (vanilla) |
| Orchestration | n8n (self-hosted) |
| Scraping | OutScraper API |
| Storage | Google Sheets API |
| Trigger | HTTP Webhook |
.
├── index.html # User interface
├── style.css # Design system (dark theme, glassmorphism)
└── script.js # Frontend logic (webhook call, progress bar, UI states)
In script.js, set your n8n webhook URL:
const WEBHOOK_URL = "https://your-n8n-instance.com/webhook/search";The n8n workflow must return a JSON response in this format:
{ "sheetUrl": "https://docs.google.com/spreadsheets/d/..." }Built by Simon — Developer, specialized in code and n8n automation.