🇮🇹 PagineGialle Italy Leads Scraper — Phone, Email, Address, P.IVA

Turn paginegialle.it — the Italian Yellow Pages — into a clean B2B lead list. Search any activity in any Italian city and get one row per business with the contact data agencies, call-centers and SaaS teams actually prospect on: name, phone, email (when published), full address with CAP and province, website, and the legal identifiers P.IVA (VAT number) / codice fiscale — plus geo coordinates, opening hours, services and ratings.

Measured on real runs, not estimated: phone fill was 100% on every sample we pulled — 75 Milano restaurants across 3 result pages, 50 Matera plumbers (mostly free listings), 25 Torino lawyers. A Milano restaurant search alone is "più di 200 risultati" at 25 businesses per page, and the Actor walks page after page for as long as you let it; PagineGialle covers every category and comune in Italy. At $3 per 1,000 leads, no login, no setup — paste an activity + a city and run.

This is the Italian sibling of our PagesJaunes France Leads Scraper — same input, same row shape, same pricing, so one pipeline can prospect both markets.

---

💡 What does this Actor do?

PagineGialle (Italiaonline) is Italy's largest public business directory. This Actor searches it like a real Chrome browser, walks the result pages (25 businesses each), opens each business's detail page, and returns a structured lead record. It beats PagineGialle's AWS WAF rate-wall automatically by rotating proxy sessions — and self-escalates to Italian residential IPs if the cheap rung ever gets challenged. You never see any of that; you just get rows.

You give it two things:

• What — an activity or category in Italian (the site's cosa box): ristoranti, idraulici, avvocati, parrucchieri, elettricisti…
• Where — an Italian city, town, or postal code (the site's dove box): milano, roma, napoli, matera, 20159…

📤 What you get — one row per business

Field	Description
name	Business / trade name
legal_name	Registered legal name from the business's structured data
phone	First phone in E.164, e.g. +390236571817 (Italian numbers keep the leading 0)
phone_display	Phone as shown on PagineGialle, e.g. 02 36571817
phones	Every displayed number (many businesses list landline + mobile)
email	Public contact email, when the business publishes one — see fill table below
address	Full address line, e.g. Via Menabrea, 33 - 20159 Milano (MI)
street / postal_code / city / province	Address split: street, 5-digit CAP, comune, 2-letter province (MI, RM…)
region	Region, e.g. Lombardia (from the detail page)
category	Activity as listed on PagineGialle
services	The services/features the business advertises (e.g. da asporto, forno a legna)
website	The business's own website
vat_number	P.IVA — the 11-digit Italian VAT number
tax_code	Codice fiscale (often = P.IVA for companies; personal format for sole traders)
rating / review_count	PagineGialle rating out of 5 + review count, when reviews exist
description	The business's own description text
opening_hours	{ "mon": ["12:00 - 15:00", "19:00 - 23:00"], … } — where shown
latitude / longitude	Geo coordinates (from the detail page)
listing_type	paid or free PagineGialle listing — free rows carry fewer optional fields
detail_url	The PagineGialle page for the business
merchant_id / scraped_at	Stable PagineGialle id + ISO scrape timestamp

Rows without at least a name + phone or address are dropped, so you never pay for empty leads.

📊 Field fill — measured on real rows, not estimated

Live samples. The middle column is the largest and most recent one (75 rows, 3 result pages, all counted by the audit re-run of 2026-08-19); the right-hand column is a small-town trade search from 2026-08-15. Sorted by fill; the sparse fields are bolded so you can't miss them.

Field	Milano ristoranti (75 rows / 3 pages, paid listings)	Matera idraulici (50 rows, 61/73 cards free)
phone / phones	100%	100%
address / city / province / postal_code	100%	100%
name / category / description / detail_url	100%	see below
latitude / longitude	99%	100%
vat_number (P.IVA)	99%	98%
legal_name / region	99%	—
opening_hours	96%	16%
services	91%	—
email	87%	14%
tax_code (codice fiscale)	79%	—
description	100%	20%
website	37%	6%
rating / review_count	36%	2%

Why website and rating are sparse — and why no scraper can fix it. Both come from the business's own structured data on its PagineGialle page: website is only present when the business has registered one (sameAs), and rating only exists once the business has at least one review (aggregateRating). We re-probed this during the 2026-08-19 audit by fetching detail pages directly: on a business with no website value, the page carries no external business link at all — the data is not hidden behind a click, it was never published. email is the same story, and it still lands at 87% on big-city paid listings.

Read the two columns before you plan an email campaign. Big-city searches are dominated by paid PagineGialle listings, which publish rich profiles — there email fill measured 80%. Small-town and trade searches are mostly free listings, which usually publish only name, address, phone and P.IVA — there email measured 14%. Every row tells you which kind it is in listing_type, and every run prints its own measured fill in the log. Phone and P.IVA are the reliable columns; email is a bonus, not a promise.

⚙️ Example input

{
  "what": "ristoranti",
  "where": "milano",
  "maxItems": 20,
  "maxPages": 1,
  "includeCompanyDetails": true
}

That block is literally the input the Actor ships with — click Try for free, hit Start, and it costs $0.06. Scale up with maxItems / maxPages when you like what you see.

Fast mode — 25 leads per request

{ "what": "avvocati", "where": "torino", "maxItems": 100, "maxPages": 4, "includeCompanyDetails": false }

With Fetch company details off, the Actor reads only the search pages — 25 businesses per HTTP request, so 100 leads ≈ 4 requests. You keep name, category, address/CAP/province and — for most listings — inline phone, description and hours. You lose email, website, P.IVA, codice fiscale, region, geo and rating (they live on the detail page). Measured on 25 Torino lawyers: phone 100%, address 100%, hours 72%, description 100%. Same price per lead either way.

📈 Example output (one real item)

A real row from an actual run (2026-08-15):

{
  "merchant_id": "878bb75c-b04a-4491-959a-0944ef863033",
  "name": "Pizzeria Ristorante L' Angolo di Napoli",
  "legal_name": "Pizzeria Ristorante L' Angolo di Napoli",
  "phone": "+390236571817",
  "phone_display": "02 36571817",
  "phones": ["02 36571817", "393 205633828"],
  "email": "ipartenopeisrl@gmail.com",
  "address": "Via Luigi Federico Menabrea, 33 - 20159 Milano (MI)",
  "street": "Via Luigi Federico Menabrea, 33",
  "postal_code": "20159",
  "city": "Milano",
  "province": "MI",
  "region": "Lombardia",
  "category": "ristoranti e trattorie",
  "services": ["da asporto", "pizzeria con forno a legna", "locali climatizzati"],
  "website": "https://www.angolodinapoli.com",
  "vat_number": "11288360966",
  "tax_code": "11288360966",
  "rating": 5,
  "review_count": 3,
  "description": "Stavi cercando un'ottima Pizza napoletana…",
  "opening_hours": { "mon": ["12:00 - 15:00", "19:00 - 23:00"], "sun": ["12:00 - 15:00", "19:00 - 23:00"] },
  "latitude": 45.49689,
  "longitude": 9.18526,
  "listing_type": "paid",
  "detail_url": "https://www.paginegialle.it/pizzeria-ristorante-l-angolo-di-napoli-milano",
  "scraped_at": "2026-08-15T04:20:51.060Z"
}

(services, description and opening_hours truncated here for length; full in the dataset.)

💰 Pricing — pay per result

$3.00 per 1,000 leads ($0.003 per lead), charged on the lead-scraped event. You are charged only for leads actually returned — nothing for empty pages, blocked requests, or duplicate cards. Rows are deduplicated on the stable PagineGialle business id before they are pushed or billed, and the run asserts uniqueness before it is allowed to succeed. The shipped default run (20 leads) costs $0.06.

🧱 How it beats the bot-wall — measured, not guessed

paginegialle.it sits behind CloudFront + AWS WAF with a rate rule: each IP gets roughly 6 requests per window, then the WAF answers HTTP 202 with a JavaScript challenge (x-amzn-waf-action: challenge), and flagged IPs briefly serve CloudFront 403. It is a rate limiter, not a hard wall — so this Actor leases each proxy session for at most 4 requests and rotates, staying under the bucket. Transport ladder, measured through Apify on 2026-08-15:

Rung	Result
Direct curl (home IP)	~6 requests, then challenged — the reason naive scrapers die
Apify datacenter, one session pushed	exactly 6 requests, then HTTP 202 challenge (4× reproduced)
Apify datacenter + rotate every ≤4 (the default)	starts every run here; holds for the whole run on some runs, gets challenged on others
Apify RESIDENTIAL + country IT (the automatic fallback)	32/32 contiguous list pages across 4 cities, 0 challenges

If datacenter sessions keep getting burned mid-run, the Actor self-escalates to RESIDENTIAL + Italy and finishes the job (unless you pinned proxy groups yourself in the Proxy input, in which case your choice is respected and no escalation happens).

How often that actually happens — measured, not assumed. Across 11 completed platform runs (2026-08-15 → 2026-08-19), the datacenter pool was challenged 0–14 times per run and the Actor escalated to RESIDENTIAL + IT on 7 of the 11, typically 5–57 s into the run. Every one of the 11 still returned its full row count — the escalation is the mechanism that makes the default reliable, not a failure. Practical consequence for your bill: expect a run to use some Apify residential bandwidth. Measured on a full 75-lead / 3-page run: $0.045 total platform usage, of which $0.030 was residential transfer — about $0.0006 per lead on top of the $0.003 per-lead price. Pin datacenter-only groups in the Proxy input if you would rather trade reliability for zero residential spend.

robots.txt: paginegialle.it does not serve one — the path returns a CloudFront 403 "Request blocked" (verified 2026-08-15), so there are no crawl-delay or disallow lines to follow. Defaults stay gentle anyway: concurrency 4, list pages carry 25 rows per request, and sessions are rotated rather than hammered.

🔁 Duplicates & pagination — measured

Contiguous page walks measured 2026-08-15: ristoranti/milano p1→p4 = 100 cards, 100 unique (0.0% dupes); p1→p8 in Roma, Napoli, Firenze and Bologna = 25 unique cards on every page, no cross-page repeats. PagineGialle does not pin promoted cards to every page the way some directories do. The Actor still dedupes on merchant_id before enriching or billing, and asserts uniqueness across everything it pushed before the run may succeed. Past-the-end pages return HTTP 404 or an empty listing — the Actor treats that as "listing exhausted", logs it, and stops cleanly.

❗ When a run fails (on purpose)

A silent empty dataset is worse than an error, so zero rows always throws, with diagnostics:

• Unknown activity term → PagineGialle refuses non-Italian / unknown terms with a persistent 403/404 (verified: a garbage term 403s while real terms on the same session return 200). The error says so and suggests Italian category names.
• Real term, zero results → the error quotes the site's own result count and reminds you the site wants Italian city names (milano, not Milan).
• Wall trouble → if every rotated session was challenged, the error says that and tells you to re-run with the Proxy input set to Apify RESIDENTIAL, country IT.
• Duplicate leak → the uniqueness assertion fails the run rather than billing a dirty dataset.

❓ FAQ

Do I need a PagineGialle account or API key? No. It reads the public directory.

Why is email sometimes empty? Because most free PagineGialle listings don't publish one (measured: 14% on a free-heavy search, 80% on a paid-heavy one). When a public email exists it is captured; otherwise the field is null. Phone is the reliable contact column at 100% measured fill.

Where do P.IVA and codice fiscale come from? From the business's own structured data on its detail page (fill measured 98–100%). Keep Fetch company details on to collect them.

Can I search a whole province or postal code? Postal codes work in Where (verified: 20159 returns the CAP's businesses). For a province, run its comuni — the row's province field lets you merge and filter afterwards.

How many results per search? 25 per page. Milano restaurants alone shows "più di 200 risultati" — raise Max result pages and Max leads to go deeper.

Does it deduplicate? Yes — by the PagineGialle business id within a run, asserted before the run can succeed. Across runs, dedupe on merchant_id or vat_number in your pipeline.

What formats can I export? JSON, CSV, Excel, HTML or an API feed — like any Apify dataset. It also works from Zapier, Make, n8n, webhooks and MCP like any other Actor.

Can I schedule it? Yes — save your input as a Task and attach an Apify Schedule for fresh leads on a cadence.

Is it fast? Yes — no browser, pure HTTP. List-only mode is 25 leads per request; detail mode adds one request per business, run concurrently.

🇫🇷🇮🇹 The family

Same input shape, same row shape, same price — one pipeline, two markets:

Actor	Market	The legal-identifier wedge
This Actor	Italy (paginegialle.it)	P.IVA + codice fiscale on ~98–100% of detailed rows
PagesJaunes France Leads Scraper	France (pagesjaunes.fr)	SIRET / SIREN / NAF on detailed rows

⚖️ Legal & fair use

This Actor collects data from a public business directory. You are responsible for how you use it and for complying with PagineGialle's / Italiaonline's Terms of Service.

GDPR note: Italian business listings can contain personal data (a sole trader's name, a direct phone, a personal codice fiscale). Under the GDPR you need a lawful basis to process personal data, must respect data-subject rights, and — for B2B prospecting in Italy — the Garante's rules on unsolicited marketing. Use the output for legitimate B2B outreach, honor opt-outs and deletion requests, and do not spam. This is guidance, not legal advice — consult your own counsel.

PagineGialle® is a trademark of Italiaonline S.p.A. This Actor is not affiliated with, endorsed by, or connected to Italiaonline.