Pricing

Rating

Developer

Actor stats

Categories

Share

Espaces Atypiques — agency team export

Who it’s for: sales, recruiters, and analysts who need a single CRM-ready table of everyone published on Espaces Atypiques agency “team” pages — no login, no API key (public website only).

What you get per person: name, job title, director vs staff band, email, phone, photo URL, long biography text, office label, and the agency page URL you can cite or revisit. Rows without an email are not exported (dataset / CSV), so you only get contactable leads.

Why one Actor: the site lists many agencies; this scraper starts from the public agency sitemap (same kind of list search engines use), fetches each French agency page, and reads the team block. You can also paste a handful of agency URLs for a quick test or a sub-region.

How to run (Apify Console)

1. Choose Sitemap (full France) or URLs (your list).
2. Optional: in Input JSON, set maxAgencies to a small number (e.g. 5) for a quick test; or pass maxResults (API) when maxAgencies is 0 to mirror the France realty hub — sitemap only. Defaults are no cap on agencies and on listing fiches in fallback mode.
3. Start the run; open the Dataset or Run log when it finishes.

Input

additionalProperties: true — you can pass extra keys from the API if you automate runs.

Example (quick test, Input JSON):

{
"mode":"sitemap",
"maxAgencies":5,
"concurrency":10
}

Example (specific agencies):

{
"mode":"urls",
"startUrls":[{"url":"https://www.espaces-atypiques.com/paris-rive-gauche/"}],
"concurrency":5
}

Output fields (main)

Role detection uses French titles (normalized): e.g. Directeur d’agence → director band; Conseiller → staff. Titles the parser does not recognise are kept in jobTitle and classified as other.

Local development

cd espaces-atypiques-agents-scraper
npminstall
echo'{"mode":"sitemap","maxAgencies":2,"concurrency":3}'> input.json
apify run --input-file=./input.json

If the CLI expects storage/key_value_stores/default/INPUT.json, use --input-file as above, or add a valid INPUT.json that matches the schema. After a run with a non-empty dataset, a semicolon CSV (UTF-8 BOM) is written to output.csv in the project root. Rows are sorted by agency path, then last name, first name, email. Columns are fixed (no dynamic Extra: fields); sourceKey is the last column; roleCategory remains on the dataset record but is omitted from the CSV to avoid overlap with band / jobTitle.

How it works (technical)

1. Sitemap mode: GET /sitemap_index.xml → find agence-sitemap.xml → GET that file → keep <loc> URLs that are not under /en/ (French pages only).
2. Fetch each agency home page (bounded concurrency, retries; redirects are followed).
3. Parse the team from .preview-collaborateur (prefer #equipe-trombi) when a mailto is present.
4. If the home page has no team block, do not call /collaborateurs/ (it often redirects to a national page). Instead, collect ventes/… URLs from the home grid (#annonces-localisations, #biens, .biens-localisation) and from the CTA that points to /ventes/?pl=… (hidden pl in the form): the Actor follows paginated search results (/ventes/page/2/?pl=…, etc.) until it has enough links or there is no next page. Then it opens each listing and parses the negotiator card (name, phone, email — including mailto in HTML comments). Total opened listings are capped by maxListingPages (default 0 = no cap; set 1–200 to limit each office — useful for smoke tests). Large full runs use more time and Apify resources. Internal safety still stops after many WordPress list pages per pl filter. listingConcurrency controls parallel listing fetches per agency.
5. Dedupe rows by sourceKey (email) and push to the default dataset. agencyPageUrl is always the sitemap office URL; dataFetchedFrom points at the listing URL when the row came from the listings fallback.

Compliance

Data is taken from public pages. You are responsible for using it in line with applicable law, the site’s terms, and GDPR (prospecting rules, contact preferences).

Support & custom work

corentin@outreacher.fr