PagesJaunes Scraper — French Business Directory (phones, ratings, SIRET)

Collect businesses from PagesJaunes.fr, the French business directory: paste a search such as plumbers in Paris and get one record per business with its name, address, activity, rating, phone numbers with PagesJaunes' direct-marketing opt-out flag, and optionally its SIRET, NAF code and websites.

Every search is checked against the number of results PagesJaunes announces ("465 résultats"): the run summary tells you whether you got all of them, and never reports a search as complete when businesses are missing.

Try it with 20

1. Open a search on pagesjaunes.fr (for example plombiers in Paris) and copy the address from your browser.
2. Paste it into Search URLs. Leave Maximum number of results at 20.
3. Start the Actor, then open Run summary before exporting the dataset.

{
  "startUrls": [{"url": "https://www.pagesjaunes.fr/annuaire/paris-75/plombiers"}],
  "maxItems": 20
}

The form prefill is 20; the default for API calls is 500. Set maxItems explicitly in API calls; 0 removes the limit.

What to paste

Input	Accepted	Notes
startUrls	https://www.pagesjaunes.fr/annuaire/<place>/<activity>	e.g. /annuaire/paris-75/plombiers, /annuaire/lyon-69/boulangeries-patisseries-artisans. Copy them from pagesjaunes.fr: the Actor never guesses a place code. A search URL of the form /annuaire/chercherlespros?quoiqui=plombiers&ou=paris-75… (what the site's sort and page links produce) is turned into /annuaire/paris-75/plombiers when ou is already a PagesJaunes place code; otherwise (e.g. ou=Paris) it is rejected with a reason. Any other URL (profile page /pros/…, page parameters on an /annuaire/… URL, other sites) is rejected with a reason, and the other URLs still run. A URL pasted twice runs once.
includeDetails	true / false (default false)	Opens each delivered business's profile page: adds siret, nafCode, websites. One more request per business, so the run is slower; the price per result does not change.
maxItems	integer, 0 = no limit	Counted across all searches. Exact: the run stops at this number, even in the middle of a results page.
proxyConfiguration	default: Apify residential proxies, France	PagesJaunes refused Apify datacenter proxies in our tests (HTTP 403, 3 of 3); residential France was accepted (2 of 2).

The Actor follows the result pages itself (20 businesses per page) and waits 2–3 seconds between requests.

Output

One record per business and per search: a business found by two of your searches appears once in each (each line is charged).

Field	Meaning
code, url	PagesJaunes establishment number (text) and its profile page.
name, address, postalCode, city	As shown in the results list. postalCode is the last 5-digit group of the address that is followed by text, city the text after it (so a postal box such as Cs 58011 … 17180 Périgny gives 17180 / Périgny).
activity	As shown, e.g. plombiers +1 (main activity + number of other activities).
rating, reviewsCount	Overall rating out of 5 and number of reviews. rating is null when the list shows no rating (never 0). The list sometimes shows a number of reviews without a rating ("19 avis"): then reviewsCount = 19 and rating = null. Both are null when the list shows neither.
phones	List of {number, optedOutOfDirectMarketing} in the site's order (some businesses have 2 or 3 numbers). [] = PagesJaunes shows no number; null = the phone lookup failed (unknown, counted in the run summary). A number not in the usual French format is kept as displayed (with a warning) rather than left out. optedOutOfDirectMarketing: true = PagesJaunes marks this number as opposed to direct marketing; null = unknown (an unrecognised element or pictogram stands next to the number — treat it as opposed: do not canvass it); false = no opt-out mark shown, which does not mean canvassing is allowed (see Personal data).
siret, nafCode, websites	With includeDetails only (otherwise null). websites = external links of the "Sites et réseaux" block, in order (internal PagesJaunes links removed).
detailsSource	profilePage when the profile page was read, else listing.
search, position, page	Which search, which results page, rank on that page. PagesJaunes reorders its results from one load to the next (in our measure, 131 of 399 positions differed 10–40 minutes later): position and page describe this run only, not a stable ranking.
scrapedAt, detailsScrapedAt	Collection time (UTC).

Real output example

Page data captured on 28 October 2026 (search paris-75/plombiers, with includeDetails). It shows the format, not today's data.

{
  "code": "06549084",
  "url": "https://www.pagesjaunes.fr/pros/06549084",
  "name": "Entreprise Bruni Sanitaire",
  "address": "125 rue Picpus 75012 Paris",
  "postalCode": "75012",
  "city": "Paris",
  "activity": "plombiers +1",
  "rating": 5.0,
  "reviewsCount": 23,
  "phones": [{"number": "01 43 43 87 11", "optedOutOfDirectMarketing": false}],
  "siret": "80478483300019",
  "nafCode": "4322A",
  "websites": ["http://www.brunisanitaire.fr"],
  "detailsSource": "profilePage",
  "search": {"what": "plombiers", "where": "paris-75"},
  "position": 18,
  "page": 2,
  "scrapedAt": "2026-10-28T15:31:38.983Z",
  "detailsScrapedAt": "2026-10-28T15:31:38.983Z"
}

Personal data

PagesJaunes lists professionals, many of whom are sole traders: their name, address and phone number are personal data under the GDPR (RGPD). You are responsible for how you use the data: you need a lawful basis, you must inform the people concerned and honour their rights (access, objection, erasure), and you must keep only what you need.

Phone numbers flagged optedOutOfDirectMarketing: true must not be used for telephone prospecting. PagesJaunes shows a pictogram on these numbers meaning the holder objects to direct marketing operations; French rules on telephone canvassing (opposition au démarchage téléphonique) do not allow calling them for prospecting. The flag is per number: a business can have one opted-out number and one that is not. In our 30-business sample (28 October 2026), 4 of 37 numbers carried it. A null flag means the Actor could not read the mark with certainty: do not canvass that number either.

false does not mean "canvassing allowed". It only means PagesJaunes shows no opt-out mark on that number. Other rules still apply (for example the Bloctel opposition list for consumers, and consent rules for SMS and e-mail); checking them is your responsibility. This Actor offers no "ready-to-call" export and does not guarantee that any use is lawful.

Run summary

Open Run summary (key-value store record SUMMARY; the same data as JSON in OUTPUT). For each search it shows:

• the status: done (businesses delivered = number announced by PagesJaunes), partial_collection (some are missing — the summary says how many and why), stopped_at_limit (maxItems reached, or the run was about to hit its timeout — see below), stopped_at_charge_limit (your maximum charge per run), not_processed, no_results (PagesJaunes says there is no result for this search), blocked_by_site, not_a_listing (PagesJaunes answered with something that is not a results list — this is not an empty search), not_found, rate_limited, http_error, network_error;
• delivered / announced, pages read;
• failure counters: phone lookups failed, profile lookups failed, businesses skipped because their listing was malformed;
• Checks: warnings such as a page without results in the middle of a search, a repeated page, or a coverage gap.

The final status message lists every non-zero failure counter; it never says a run went fine when something failed. The run fails only when no search delivered anything ("No business found, check the search URLs").

Run timeout. When less than about 2 minutes remain before the run's timeout, the Actor stops cleanly: everything already collected stays in the dataset, the search in progress gets stopped_at_limit ("run time limit"), the next ones not_processed, and the Run summary is written. Raise the timeout and start again for the rest.

OUTPUT.bytesReceived counts the bytes of the responses received (decompressed bodies): an approximation of the proxy traffic, which is usually lower (compressed) plus headers.

Pricing

Pay per result: $1.50 per 1,000 records (one event per business delivered; nothing is charged for rejected URLs or skipped businesses), plus a $0.005 start fee per run (once per run up to 1 GB of memory; the default is 256 MB). Apify subscribers pay less automatically, on both events: 10% off on the Bronze plan, 20% on Silver, 30% on Gold and above.

Records	Price
20	$0.035
1,000	$1.505

includeDetails does not change the price, but makes the run longer (one more request per business, 2–3 s apart). Measured on the platform on 28 October 2026: 402 records with details in about 40 minutes, i.e. roughly 1 h 40 min per 1,000. Raise the run timeout for large runs. The Actor needs little memory (default 256 MB: no browser).

Known limits

• One phone request per delivered business (PagesJaunes loads phone numbers on click), plus one profile request with includeDetails: large runs are slow by design (2–3 s between requests).
• The search URL must be copied from pagesjaunes.fr: the Actor does not build URLs from free text, to avoid guessing a place.
• Coverage is compared with the number PagesJaunes announces; if PagesJaunes stops serving pages before that number, the search is reported as partial_collection.
• rating comes from the results list. A profile page can show a rating the list does not (pilot sample: rating on 20 of 30 profile pages, 14 of 30 in the list); the record follows the list.
• websites keeps external links only; links to other PagesJaunes pages are removed.
• Opening hours, photos, descriptions and reviews text are not collected.
• Behaviour after a platform migration: the Actor resumes without duplicates by reading the search again from page 1 and skipping businesses already delivered.

Support

Open an issue on the Actor's Issues tab with the run ID and the search URL.

This Actor is an independent tool. It is not affiliated with, endorsed by or sponsored by PagesJaunes or Solocal; "PagesJaunes" is a trademark of its owner, used only to describe the data source.

En français

PagesJaunes Scraper récupère les professionnels d'une recherche PagesJaunes (« plombiers à Paris ») : nom, adresse, activité, note, numéros de téléphone avec le pictogramme d'opposition au démarchage, et en option SIRET, code NAF et sites web (includeDetails).

• À coller : l'adresse d'une recherche copiée depuis pagesjaunes.fr, de la forme https://www.pagesjaunes.fr/annuaire/paris-75/plombiers. Les autres adresses sont refusées avec une raison, sans bloquer les autres.
• Essai : maxItems = 20 (préréglé dans le formulaire). Lisez ensuite le Run summary : il compare le nombre d'établissements livrés au nombre annoncé par PagesJaunes (« 465 résultats ») et ne dit jamais « terminé » s'il en manque.
• Téléphones : phones est une liste dans l'ordre du site ; [] = aucun numéro affiché ; null = la requête a échoué (inconnu, compté dans le résumé).
• Opposition au démarchage : un numéro marqué optedOutOfDirectMarketing: true porte le pictogramme PagesJaunes d'opposition aux opérations de marketing direct. Ces numéros ne doivent pas être démarchés par téléphone (opposition au démarchage téléphonique, droit français). null = marque illisible avec certitude : ne pas démarcher non plus. false ne veut pas dire « démarchage autorisé » : d'autres règles s'appliquent (Bloctel pour les particuliers, consentement pour SMS et e-mail). Vous êtes responsable de l'usage des données (RGPD) : base légale, information des personnes, respect de leurs droits.
• Prix : 1,50 $ pour 1 000 résultats, détails compris, plus 0,005 $ de démarrage par run (0,035 $ pour 20). Remise automatique pour les abonnés Apify : −10 % (Bronze), −20 % (Silver), −30 % (Gold et plus).
• Outil indépendant, sans lien avec PagesJaunes ni Solocal.