Welcome to the Jungle Jobs Scraper

Get job offers from welcometothejungle.com as clean, typed records: title, company, locations, contract, remote policy, salary with its period and currency, publication date, missions, benefits, and optionally the full job page (description, profile, skills, tools, address, apply link, status).

It replays the request the website's own company and SEO pages send to their public job index (same index, same filters, same page size), so the records follow the jobs and the order of the page you pasted. Searches with more than the 1,000 results the site serves per query are split automatically, and the run summary tells you, per URL, how many jobs the site announced and how many you got.

Try it with 20 jobs

1. Keep the prefilled start URL (https://www.welcometothejungle.com/fr/companies/sii/jobs) and Maximum number of jobs = 20.
2. Click Start. The run takes a few seconds and costs $0.015 (or $0.035 with job details), start fee included.
3. Open the Jobs tab, then Run summary to see what was announced and what was collected.

What to paste

You paste	Example	What you get
A company jobs page, with the filters you clicked on the site	https://www.welcometothejungle.com/fr/companies/thales/jobs?refinementList%5Bcontract_type%5D%5B%5D=internship	Every job of that page, all pages (30 per page), filters kept exactly
An SEO jobs page	https://www.welcometothejungle.com/fr/pages/emploi-paris-75000	The jobs that page lists (20 per page), all pages
A single job page	https://www.welcometothejungle.com/fr/companies/sii/jobs/developpeur-java-confirme-f-h_toulouse_GS_mrpQZbV	That job, with its details; a job that is no longer open comes with status: "archived"
/en/ pages	https://www.welcometothejungle.com/en/companies/mistral-ai/jobs	Same jobs, English labels and URLs

• Group pages. Some company pages list the jobs of a whole group: /fr/companies/edf/jobs shows the jobs of EDF and 9 subsidiaries (Framatome, EDF Renouvelables…), 58 apprenticeships where "edf" alone has 51. The Actor reads which companies the page lists from the site's own page configuration (one small request per company page, cached for the run) and filters exactly like the page, so you get the same jobs, the same total and the same page/position. If that configuration cannot be read, it filters on the company of the URL alone, warns you (COMPANY_FILTER_FALLBACK) and marks the URL partial_collection ("company filter could not be read: subsidiaries of a group may be missing"), never done: the total it compares to is then the company's own, not the page's. A temporary error is retried once when the run resumes after a migration.
• Company page that does not exist. Some companies have jobs on the site but no jobs page (e.g. /en/companies/l-oreal/jobs answers "Error 404"). Their jobs are still in the site's index and their job pages are public: the Actor delivers them, filtered on the company of the URL, with the warning COMPANY_JOBS_PAGE_NOT_FOUND.
• page= in a pasted URL is ignored: collection always starts at page 1.
• Keyword search is not public on the site any more. Since October 2026 the site shows keyword search results only to logged-in users (/fr/jobs?query=... displays a sign-up page). Old keyword URLs are still accepted and translated to the job index, with a clear warning; the keyword field does the same. Because there is no public page to compare with, we do not promise "same as the website search" for those.
• Keyword mode filters: countryCode (FR by default, empty = all), contractTypes, remote, city, language (fr/en labels and URLs), postedSince (YYYY-MM-DD).

Output

One record per job and run (duplicates across start URLs are delivered once). Unknown values are null: never an empty string, an invented 0, false or "no". A real record (texts shortened with "…"):

{
 "jobId": "4a986e90-1737-4247-a41e-f6820e8d6e81",
 "jobReference": "GS_A649137",
 "url": "https://www.welcometothejungle.com/fr/companies/sii/jobs/ingenieur-ivv-logiciel-embarque-aeronautique-f-h_toulouse_GS_A649137",
 "title": "Ingénieur(e) IVV Logiciel Embarqué Aéronautique (F/H)",
 "company": "Groupe SII", "companyId": "5p9XmWL", "companySlug": "sii",
 "companyUrl": "https://www.welcometothejungle.com/fr/companies/sii",
 "companyLogo": "https://cdn-images.welcometothejungle.com/-OCGJ1AoQbY-MO_H0BCX9Dq_3njRpPrSEcpoBgJmud8/rs:a…",
 "companySize": 16000,
 "location": "Toulouse", "locations": ["Toulouse"],
 "offices": [{"city": "Toulouse", "localCity": "Toulouse", "district": "Haute-Garonne", "state": "Occitanie", "country": "France", "countryCode": "FR"}],
 "contractType": "full_time", "contractTypeLabel": "CDI",
 "contractDurationMin": null, "contractDurationMax": null,
 "remote": "no", "remoteLabel": "Télétravail non autorisé",
 "salaryMin": 35000, "salaryMax": 60000, "salaryPeriod": "year", "salaryCurrency": "EUR",
 "experienceLevelMin": null, "educationLevel": null,
 "postedAt": "2026-10-28T23:03:24Z",
 "summary": "Rejoignez SII, un leader mondial de l'industrie aéronautique. En tant qu'Ingénieur Vérific…",
 "keyMissions": ["Développement de scripts de test et mise en œuvre des cas de test auto…", "…"],
 "benefits": ["Horaires de travail flexibles", "Entre 1-2 jours de télétravail", "…"],
 "sectors": ["Logiciels", "Intelligence artificielle / Machine Learning", "IT / Digital"],
 "language": "fr", "isBoosted": false,
 "sourceUrl": "https://www.welcometothejungle.com/fr/companies/sii/jobs",
 "page": 2, "position": 6, "searchPart": null,
 "scrapedAt": "2026-10-29T12:06:50+00:00",
 "detailsSource": "jobPage",
 "descriptionHtml": "<p>Dans le cadre d&#39;une mission stratégique pour un leader mondial de l&#39;industrie a…",
 "descriptionText": "Dans le cadre d'une mission stratégique pour un leader mondial de l'industrie aéronautique…",
 "profileHtml": "<p>Compétences requises :</p><p>C, Doors, Jira, Python</p><p>Qualités désirées :</p><p>For…",
 "profileText": "Compétences requises :\n\nC, Doors, Jira, Python\n\nQualités désirées :\n\nForce de proposition,…",
 "recruitmentProcess": null,
 "applyUrl": "https://sii-group.com/fr-FR/offers/ingenieure-ivv-logiciel-embarque-aeronautique-fh?source…",
 "skills": ["Communication", "Gestion de la conformité"], "tools": ["Linux", "Python", "JIRA"],
 "startDate": null, "address": "Toulouse, Occitanie, France", "zipCode": "31000",
 "latitude": 43.60579, "longitude": 1.44863,
 "profession": "Ingénierie matérielle", "professionCategory": "Technologie et ingénierie",
 "updatedAt": "2026-10-28T23:03:24.635880Z", "status": "published",
 "jobPageSalaryMin": 35000, "jobPageSalaryMax": 60000, "jobPageSalaryPeriod": "year", "jobPageSalaryCurrency": "EUR"
}

Good to know:

• remote keeps the site's code: no, partial, punctual, fulltime, or unknown (about 45% of the index: the card shows no remote label). unknown is never turned into "no"; its remoteLabel is null.
• Salary: salaryMin alone means "from" (the card shows "≥ 55K €"), salaryMax alone means "under" ("< 300 € par jour"); salaryPeriod is year, month or day. When the site stores a period without any amount, all four salary fields are null (nothing is displayed). Amounts are numbers (45000, not 45000.0 or "45K"). An amount of 0 in the index is delivered as null (the card shows no salary and the job page says "Non spécifié").
• Two salary sources, kept apart. salaryMin/salaryMax/salaryPeriod/salaryCurrency come from the search index, exactly what the job card shows. With includeDetails, jobPageSalaryMin/jobPageSalaryMax/ jobPageSalaryPeriod/jobPageSalaryCurrency hold the salary shown on the job page itself. They usually agree, but the site does not always: one job had no salary on its card and "35K à 40K €" on its page. Use the job page values when you want the most complete salary, the index values when you want what the listing shows.
• locations uses the local city name the card shows ("Genève", not "Geneva") in the card's order; location is the full list joined by ", " even when the card truncates it ("… +18").
• language is the language of the job ad (fr, en, es…), not the language of the page you pasted.
• jobId is the site's UUID; jobReference (the short code at the end of most job URLs) is null for about 16% of jobs.
• contractTypeLabel / remoteLabel are the labels displayed in the language of your URL; a code whose label was not measured gets null (the code is still delivered).
• searchPart is null for jobs read from the site's own pages; for jobs found by splitting a search over 1,000 results, it holds the extra filter of that part (e.g. contract_type:internship).
• Detail fields (descriptionHtml … status) are filled only with includeDetails (or for a pasted job page); otherwise they are null and detailsSource is listing.

Run summary

Each run writes a readable Run summary (key-value store record SUMMARY) and the same data as JSON (OUTPUT): per start URL, its status, the jobs delivered, the jobs the site announced (jobsAnnouncedBySite), the distinct jobs seen (jobsSeen), what is missing and why.

Status	Meaning
done	Every announced job was seen and delivered
no_new_jobs	Every job of this URL was already delivered by an earlier start URL
stopped_at_limit	Your maxItems or maximum charge was reached
partial_collection	Fewer distinct jobs than announced, with the reason (page that failed, empty page, page that only repeated jobs, page count that dropped, split part that failed). Never reported as done
no_results	The site announced 0 jobs: the search is empty, the URL is wrong or the site is down (warning NO_RESULTS). A run where every URL is empty fails
job_not_available	A pasted job URL served the site's "page not found" (sent with HTTP 200) or another job; nothing delivered or charged
blocked_by_site, http_error, network_error, unrecognized_page	The site refused or failed; this is never reported as an empty search

Warnings you may see: ANO_COUVERTURE (partial collection, with the numbers), SEARCH_URL_NOT_PUBLIC, COMPANY_JOBS_PAGE_NOT_FOUND, COMPANY_FILTER_FALLBACK, JOB_ARCHIVED, ANO_DETAILS / DETAILS_STOPPED (job pages not read), UNEXPECTED_DATA / UNKNOWN_VALUE (one odd value: the job is delivered without that field, the run continues), KEY_FALLBACK, PROXY_SETTING. The final status message never says everything went fine when one of these counters is not zero. When a job page answers something other than a normal page (for example an HTTP 202 anti-bot challenge), the first 3 such answers are saved in the key-value store record DEBUG_PAGES (status, response headers, first 3,000 characters, final URL, proxy session number; no password or key) so the cause can be checked.

Pricing

Pay per result, no subscription:

• $0.0005 per job delivered (apify-default-dataset-item);
• + $0.001 per job whose page was read for details (job-details, only with includeDetails or a pasted job page);
• + $0.005 per run to start (apify-actor-start, once per GB of memory; the default is 1 GB).

Apify subscribers pay less automatically on jobs and on the run start: 10% off on the Bronze plan, 20% on Silver, 30% on Gold and above. The job-details charge has no discount: $0.001 per job page on every plan.

	20 jobs	1,000 jobs
Without details	$0.015	$0.505
With details	$0.035	$1.505

These totals include the $0.005 start fee.

The job-details price includes the residential proxy traffic that job pages need (about 85-92 KB transferred per page, measured). A job whose page could not be read is never charged the details event. Your maximum charge per run is respected: the run stops cleanly before reading a job page it could not pay for (a pasted job page is only read when the job and its details both fit in your maximum charge).

Known limits

• Keyword search is no longer public on the site (see "What to paste"); keyword results come from the same index but cannot be compared with a public page.
• The site serves at most 1,000 results per query. Larger searches are split by contract type, remote policy, then publication date; jobs published or removed while the run splits a large search can make the count differ, and the summary says so (partial_collection, never done).
• Filters other than refinementList[...] and query in a pasted URL (for example a sort) are not applied; the summary lists them (IGNORED_URL_PARAMETERS).
• The site's company page counter includes a "Candidature spontanée" card (e.g. 479 for 478 jobs); it is not a job and is neither counted nor delivered.
• A job that is no longer open is not in the index: it is only reachable as a pasted job page, delivered with status: "archived" and the fields its page gives (sectors, size, logo and boost stay null). For such a job the salary fields come from its job page, the only source left (they then equal jobPageSalary*); for every other job salary* is the index value the card shows.
• Bot challenge on some job pages. On Apify, about 5% of job pages answer with an anti-bot challenge (HTTP 202 from CloudFront, x-amzn-waf-action: challenge, AWS WAF) instead of the page, even after new IPs. A browser solves it, an HTTP client cannot. Those jobs are delivered without details and without the job-details charge; the run summary counts them, and the first answers are kept in DEBUG_PAGES.
• The job page needs residential proxy. The search index works from datacenter IPs, but the site refuses job pages from datacenter IPs (HTTP 403, measured 2026-10-29). Job pages and SEO pages therefore use their own setting, Proxy for job pages (detailsProxyConfiguration, Apify RESIDENTIAL, country FR, by default); a refused page is retried with a new IP, up to 5 times. Proxy for the search index (proxyConfiguration) only applies to the index. If you set the job pages proxy without the RESIDENTIAL group, or turn it off, the run does as you ask and warns you (PROXY_SETTING): details will likely fail.
• Run time (measured on Apify). Without details, the index is fast: 7,139 jobs ("developpeur", France) in about 1 minute. With details, count about 1 second per job (one job page each, about 288 KB uncompressed and 85-92 KB transferred): 16-20 minutes per 1,000 jobs. The default run timeout is 4 hours (about 14,000 jobs with details); for a larger collection with details, raise the timeout in the run options, otherwise the run stops before writing its summary.

Support

Found a job the site shows that the run missed, or a field that differs from the page? Open an issue on the Actor's Issues tab with the start URL and the run ID; the run summary usually points to the cause.

This Actor is not affiliated with, endorsed or sponsored by Welcome to the Jungle. "Welcome to the Jungle" is a trademark of its owner. It reads publicly available job offers only; use the data in line with the site's terms and the law that applies to you (GDPR for personal data).

En français

Collez une page entreprise de welcometothejungle.com avec les filtres choisis sur le site (stage, alternance, télétravail…), une page SEO (/fr/pages/emploi-paris-75000) ou une fiche d'offre : vous obtenez une ligne par offre (titre, entreprise, lieux, contrat, télétravail, salaire avec sa période, date, missions, avantages ; avec les détails : description, profil, compétences, outils, adresse, lien de candidature, statut). La recherche par mot-clé n'est plus visible sans compte sur le site : elle reste possible ici, avec un avertissement. Au-delà de 1 000 résultats, la recherche est découpée automatiquement ; le résumé du run dit, URL par URL, combien d'offres le site annonçait et combien ont été livrées. Prix : 0,50 $ les 1 000 offres, 1,50 $ avec les détails, plus 0,005 $ de démarrage par run ; remise automatique pour les abonnés Apify sur les offres et le démarrage (−10 % Bronze, −20 % Silver, −30 % Gold et plus) ; pas de remise sur les détails (0,001 $ par fiche quel que soit l'abonnement).