Job Feed: 110+ Job Boards & Career Pages in 60+ Countries
Pricing
$0.50 / 1,000 unique job in aggregated feeds
One deduplicated, newest-first job feed from 110+ job boards in 60+ countries, 40+ ATS career systems and any company career page. Deepest in Switzerland, Germany and Austria. Daily new-jobs-only runs, radius filter, agency flag, and MCP for AI agents.
_type
Optional
Item kind: 'job' (default, omitted for jobs). By default the dataset holds ONLY job rows -- run_summary/cost_report/relevance_report/sources_manifest live in the run's OUTPUT key-value-store record instead (see output_schema.json's runManifest field). Set input's includeReportRowsInDataset=true to also push these four as extra rows here, same as job-feed did before 0.9.6x. 'company_hiring' is the only row kind when outputMode is "companies" (one per companyUrls entry; see the hiring_* fields below). 'company_no_matching_jobs' is always a real dataset row (not gated behind includeReportRowsInDataset): a company whose career site was read successfully but had zero jobs matching your filters this run -- see jobs_on_site/jobs_matching/reason/note/filters below. Never charged. Set input's includeCompaniesWithoutJobs=false to stop getting these and receive job rows only; filter client-side on _type either way (e.g. _type == null for jobs only, since there is no dataset-schema view for row-level filtering).
_platform
Optional
First-arriver platform for this job row. The platform whose values are preserved when multiple sources report the same job_id. Complete per-source list lives in sources[].
job_id
Optional
Stable canonical hash of the job across sources and across runs. First 16 hex chars of SHA-256 over folded title + company + city.
job_id_aliases
Optional
job_id plus the own ids of the rows merged into this posting when it was pushed (sorted, unique). The same posting can come back under another job_id on a later run, because job_id is the first arriving row's id and company-name variants hash differently. Check [job_id, ...job_id_aliases] against your seen-set. Rows merged after this one was pushed are listed in the sources_manifest record under additional_job_id_aliases_by_job_id.
title
Optional
Job title (HTML stripped, UI chrome cleaned)
company
Optional
On a _type: company_hiring row (outputMode "companies") this is the company label as sent (string). Otherwise: Company block. Also carries ref (the companyUrls ref you sent, omitted when none). Always populated: name, url (external website), domain, logo_url. Populated when the sub-actor emits them (usually only under includeCompanyDetails=true): about, industry, size_range, phone, email, headquarters, address {street,city,region,postcode,country}, profile_url (FIRST-ARRIVER platform's company profile page — for the full per-source list iterate sources[].company_profile_url), career_page_url (company's own careers portal), ats_detected (greenhouse|lever|workday|ashby|personio|smartrecruiters|breezy|jobvite|icims|taleo|successfactors|teamtailor|recruitee|bamboohr|softgarden|workable|pinpoint), social_urls {linkedin,facebook,twitter,instagram,youtube,tiktok,xing,github,kununu,glassdoor}, active_jobs_count, active_jobs[] (up to 20 {title,url} at same company), recruiters[] (company-level roster; dice only: name, title, email, profile_url), rating (*float64, platform-specific 0-5 scale; ~11 actors), review_count (*int, number of reviews behind rating), founding_year (*int, 4-digit year; linkedin-go/glassdoor/builtin/levels-fyi/etc.), benefits ([]string; ~20 actors: jobs-ch-go, arbeitsagentur-de, jobware-de, karriere-at-go; alias company_perks for US actors), employee_count_from (*int, lower bound of employee headcount range parsed from company_employees; distinct from size_range which holds the raw string), employee_count_to (int, upper bound; nil for open-ended ranges like 1000+), platform_extras (opaque passthrough for platform-specific company_ fields not covered above). On a _type: company_no_matching_jobs row this is instead the smaller CompanyNoMatchCompany shape: name, domain, url, career_page_url, ats_detected only. Agency flags (always present): is_agency (true for a DACH-registry staffing, executive search, RPO, job-aggregator or engineering-services firm with high or medium confidence, or a name keyword such as Zeitarbeit or Personalvermittlung), agency_type (staffing|executive_search|rpo|job_aggregator|engineering_services|placeholder, or null), agency_confidence (high|medium|licence_only|keyword, or null), is_agency_signal (true for any registry hit including licence_only, or a keyword hit). licence_only means the firm only holds a staffing licence, often as a side business: a signal, not a verdict.
location
Optional
Location block: raw, city, region, country, remote (onsite|hybrid|remote), postal_code (job's own postal/zip code; aliases post_code, postal_code, zip, zip_code, postcode; omitted when no sub-actor field is present)
employment_type
Optional
permanent | temporary | contract | internship | apprenticeship | null
workload_pct
Optional
Workload percentage: {min, max}. Mandatory for CH.
salary
Optional
Salary block: {min, max, currency, period, text, source}. Never currency-converted. text holds the raw human-readable string (e.g. '50k-70k gross yearly') when min/max can't be parsed. source is 'stated' (the posting's own salary) or 'hidden_field' (a band recovered for an ad that hides its salary), null when there is no salary data.
posted_date
Optional
ISO8601 date the listing was posted
expires_at
Optional
ISO8601 listing expiration / application deadline. ~45 actors. Aliases consumed in priority order: expires_at, listing_expires_at, valid_through, closing_at, closing_date, valid_until.
first_seen
Optional
ISO8601 timestamp within THIS run only. Cross-run history is in lifecycle (trackJobLifecycle).
last_seen
Optional
ISO8601 timestamp within THIS run only.
lifecycle
Optional
(includeReportRowsInDataset=true only, and only when trackJobLifecycle is on) Run summary: enabled, store_name, store_size, tracked_this_run, closed_detection (ran, skipped, skipped_incomplete or first_run), closed_count, closed_detection_sources_skipped
content_hash
Optional
SHA-256 over mutable fields (description, employment_type, salary, workload_pct) — for detecting posting changes across your runs
language
Optional
Detected listing language
description_text
Optional
Full listing text (HTML stripped). Present only if includeJobDetails=true and source supports it.
description_html
Optional
Raw HTML description when the sub-actor emits it (~35 platforms). Empty otherwise.
description_md
Optional
Markdown description when the sub-actor emits it (~25 platforms). Empty otherwise.
apply_url
Optional
Apply URL (may be the source listing or an external ATS)
contacts
Optional
Per-job recruiter contacts (salutation, first_name, last_name, email, phone, position, raw, source_platform). source_platform is the id of the board the contact came from and is kept when duplicate rows are merged. Populated when includeContactData=true and the sub-actor emits contact fields (contact_* keys, contact arrays, and the board-specific keys apply_email, job_contact_email, job_contact_phone, contact_person, job_poster_name, job_poster_title, contact_title, contact_department, craigslist_reply_email).
job_status
Optional
Canonical posting status: open, expired or unknown. Mapped from the sub-actor's job_status (online/offline/expired and similar), is_active, expired or is_expired; unknown when the board reports none. A row existing or an HTTP 200 never means open on its own.
job_url_status
Optional
Health of job_url when the board reports it (career-site-jobs Abacus): advertisement_off_apply_url = job_url is the application page and works; portal_frontend_off = neither the ad nor the apply page opens. Other values pass through verbatim.
requirements
Optional
Structured list of job requirements / qualifications. ~30 actors: DACH Go family, bayt, karriere-at-go, anofm-ro, hh-ru. Accepts []string or single newline-delimited string from sub-actor.
skills
Optional
Structured list of required/desired skills. ~20 actors: stepstone-de-go, dice, linkedin-go, profession-hu, cwjobs-uk, hellowork-fr, naukri-scraper, trudvsem-ru, rigzone-us, jobserve-uk, totaljobs, indeed.
department
Optional
Internal organizational department or team name. career-site-jobs-scraper, career-site-scraper, enterprise-ats-jobs-scraper, higheredjobs, usajobs, ge-ch-emploi, jobs-ac-uk, infojobs.
experience_level
Optional
Seniority tier as emitted by the source platform (e.g. 'senior', 'mid-level', 'entry'). ~25 actors. Alias: seniority_level (linkedin-go).
education_level
Optional
Minimum academic qualification required. ~15 actors: DACH Go family, anofm-ro.
industry
Optional
Job-level industry / sector tag. Distinct from company.industry (the company's own declared sector). ~13 actors: anofm-ro, finn-no, hh-ru, escape-the-city, infojobs, web3career, work-ua, workatastartup, ziprecruiter, wayup-us, cryptojobslist, dcz-gov-ua, trudvsem-ru.
technologies
Optional
Tech stack tags as structured arrays. ~6 actors: swissdevjobs-ch-go, workatastartup, nofluffjobs-eu, justjoinit, welcome-to-the-jungle. Aliases: technologies, tech_stack. Empty list [] when unavailable.
ats_system
Optional
The ATS/job-system platform that served this job (e.g. workday, greenhouse, phenom). Populated for company career-page rows only: website-job-extractor{,-browser} pass their own ats_system through verbatim; career-site-jobs reports the platform it extracted from. null for board-platform rows (indeed, linkedin, etc.), which never have a company job system to begin with.
ats_url
Optional
URL of the ATS/job-system listing page the job was read from. website-job-extractor{,-browser}: their own ats_url, verbatim. career-site-jobs: a best-effort match against the URL you dispatched to it (null when no confident match is possible). Feed this back as a careerSites entry on your next run to skip straight to career-site-jobs instead of back through the extractor -- see the OUTPUT record's next_run_input.
ats_tenant
Optional
Tenant identifier within the ATS (e.g. the Workday tenant slug). website-job-extractor{,-browser} only, verbatim. Always null for career-site-jobs and board-platform rows.
career_page_url
Optional
The career page URL where job discovery started, as reported by website-job-extractor{,-browser}. null for career-site-jobs and board-platform rows. Distinct from company.career_page_url, which is derived for every platform from a best-effort ATS guess.
matched_query
Optional
The specific keyword/search-query string that surfaced this result at the sub-actor level (when the sub-actor emits search_query).
relevance_score
Optional
Fraction of user-supplied keyword tokens present in the title (0.0-1.0). 1.0 when no keywords were provided.
relevance_matched_tokens
Optional
The specific keyword tokens from the user's query that matched the title. Empty when no keywords were provided or none matched.
sources
Optional
Every platform this job was found on. Array by design — a role posted on 3 boards keeps 3 entries. Each SourceRef carries: platform, url (source listing URL), source_id, first_seen, company_profile_url (this platform's OWN company profile page), company_name (raw casing/legal-form suffix as this platform emitted it). In streaming dedup mode this array has length 1 (first-arriver only) and the other sources live in the sources_manifest terminal item. In batch dedup mode it has length N with all sources inline.
hiring_organization
Optional
The employing entity when it differs from the company the board belongs to, for example a group's subsidiary (company Johnson & Johnson, hiring_organization Cilag AG). career-site-jobs and any other sub-actor that reports it. Absent when not reported.
application_deadline
Optional
Application deadline as stated in the posting (ISO date when parsed). From website-job-extractor with includeJobDetails. Also used as expires_at when no expiry is given. Absent when not reported.
reference_number
Optional
The employer's own reference or requisition number for the posting. Absent when not reported.
apply_url_is_synthetic
Optional
True when apply_url was built by the extractor as a page#job-slug anchor because the listing has no posting page of its own. Absent when false.
is_easy_apply
Optional
True when the posting can be applied to on the source site itself. Absent when not reported.
street_address
Optional
Street address of the workplace from the posting detail page. Absent when not reported.
working_hours
Optional
Working hours as written in the posting, for example Mo-Fr 08:00-17:00. Absent when not reported.
updated_at
Optional
Date the source last refreshed or re-published the ad, as the sub-actor reports it (jobbasel-ch: ISO 8601). posted_date stays the first publication date. Absent when not reported.
start_date
Optional
Fixed employment start date (arbeit-swiss: ISO date). Absent when not reported.
starts_immediately
Optional
True when the ad says to start immediately (arbeit-swiss). Absent when not reported.
end_date
Optional
End date of a fixed-term job (arbeit-swiss: ISO date). Absent when open-ended or not reported.
work_forms
Optional
Special work forms such as night work, shift work, sunday and holiday work, home work (arbeit-swiss). An empty array means the ad has none; absent means the source did not report it.
locations
Optional
Every city of a multi-location ad in page order (stepstone-de, stellenanzeigen-de). Absent when not reported.
hours_per_week
Optional
Weekly hours stated in the ad as {min, max}; a single value has min equal to max (stellenanzeigen-de). The text as written is in working_hours; workload_pct stays the percent range. Absent when not reported.
languages
Optional
Languages the posting asks for. Absent when not reported.
input_refs
Optional
Rows read from a careerSites or companyUrls entry only: the input entries that produced the row, as {source: "careerSites"|"companyUrls", index: 0-based position in that input array, url, company, employer} with url/company/employer verbatim from your input (null when not given). An array because the same job found through two entries lists both. Use it, not company.name (the employer's display name), to attribute rows to your input. Absent on board and keyword rows.
platforms_requested
Optional
(includeReportRowsInDataset=true only) Run summary: platform count requested
platforms_succeeded
Optional
(includeReportRowsInDataset=true only) Run summary: platform count that returned data
platforms_failed
Optional
(includeReportRowsInDataset=true only) Run summary: platforms that errored or timed out
platforms_skipped_degraded
Optional
(includeReportRowsInDataset=true only) Run summary: degraded/unvalidated platforms excluded from defaults
raw_results
Optional
(includeReportRowsInDataset=true only) Run summary: pre-dedup result count
unique_jobs
Optional
(includeReportRowsInDataset=true only) Run summary: post-dedup unique job count
duplicates_merged
Optional
(includeReportRowsInDataset=true only) Run summary: cross-source merge count
only_new
Optional
(includeReportRowsInDataset=true only, and only when onlyNewJobs is on) Run summary: seen_skipped, new_pushed, store_name, store_size, ttl_days
orchestrator_billed_events
Optional
(includeReportRowsInDataset=true only) Run summary: unique jobs billed at orchestrator rate
total_cost_usd
Optional
(includeReportRowsInDataset=true only) Run summary: sum of sub-actor cost + orchestrator cost
per_platform
Optional
(includeReportRowsInDataset=true only) Cost report: per-platform cost + unique contribution
subactor_cost_total_usd
Optional
(includeReportRowsInDataset=true only) Cost report: sum of all sub-actor charges this run
orchestrator_cost_usd
Optional
(includeReportRowsInDataset=true only) Cost report: this actor's own event cost
cost_per_unique_job_usd
Optional
(includeReportRowsInDataset=true only) Cost report: total_cost / unique_jobs
overlap_matrix
Optional
(includeReportRowsInDataset=true only) Cost report: pairwise platform overlap statistics
redundancy_hint
Optional
(includeReportRowsInDataset=true only) Cost report: human-readable suggestion for which platform to drop
ref
Optional
_type: company_hiring rows only (outputMode "companies"). The ref you sent on this companyUrls entry (for example a Google Maps place_id), echoed unchanged; the merge key. null when none was sent. On job rows, company.ref carries the same value.
website
Optional
_type: company_hiring rows only (outputMode "companies"). The website URL as sent (after trimming and adding https:// to a bare domain).
hiring_checked
Optional
_type: company_hiring rows only (outputMode "companies"). false only when the company could not be read (hiring_status not_read).
hiring_status
Optional
_type: company_hiring rows only (outputMode "companies"). hiring (at least one open role), no_open_roles (the site was read, nothing is open), or not_read (the site could not be read, or the entry is a duplicate or invalid; see not_read_reason).
hiring_open_roles_count
Optional
_type: company_hiring rows only (outputMode "companies"). The site-wide number of open roles when the reader reports a total for this request, otherwise the number of distinct roles read. null for not_read. A country, keyword, location, posted-within or title filter makes the count the roles read, not the site total.
hiring_open_roles_count_is_capped
Optional
_type: company_hiring rows only (outputMode "companies"). true only when no site total is known and the roles read equal maxResultsPerCompany, so more roles may exist.
hiring_roles
Optional
_type: company_hiring rows only (outputMode "companies"). Titles of up to maxResultsPerCompany open roles.
hiring_roles_detail
Optional
_type: company_hiring rows only (outputMode "companies"). Aligned with hiring_roles (same order, same length, at most 50): objects {title, location_raw, city, postal_code, country}, null for unknown parts. Empty array for no_open_roles and not_read.
hiring_evidence
Optional
_type: company_hiring rows only (outputMode "companies"). How the status was established. hiring: ats_board (roles from an ATS or career-site-jobs), careers_page (from the company website) or job_board (only a job board); the first applicable in that order. no_open_roles: ats_zero (a job system was read, 0 jobs), careers_page_empty (a real careers page was read, 0 jobs) or no_careers_page (the homepage was fetched fine and has no careers, jobs or job-system link). null for not_read.
hiring_sources
Optional
_type: company_hiring rows only (outputMode "companies"). Where the roles were read: ATS names (greenhouse, lever, workday, ...) or website.
hiring_jobs_url
Optional
_type: company_hiring rows only (outputMode "companies"). The careers page or ATS board URL, or null.
hiring_checked_at
Optional
_type: company_hiring rows only (outputMode "companies"). ISO 8601 time the row was built.
not_read_reason
Optional
_type: company_hiring rows only (outputMode "companies"). Why the company was not read (blocked, board_not_found, sub_run_failed, duplicate_url, invalid_url, directory_profile_url (the website is a directory, social or profile site and was not read), no_result, needs_browser, charge_limit (the source was not started because your max charge was used up), dns_error (the website domain does not resolve), ats_handover_available, ...), else null. ats_handover_available: the site uses a job system we could not read in this run.
jobs_on_site
Optional
_type: company_no_matching_jobs rows only. Total job count the career site itself reported, BEFORE your keywords/location filter was applied -- only ever set when that total is actually known independent of the filter. 0 means a genuinely empty career page; >0 means jobs exist there but none matched your filters this run. null for a career-site-jobs row whenever its ATS's own search already applied your filter server-side (its ATS counted 0 results for your exact search) -- any count that ATS reports in that mode is itself a result of the filter, not an independent site-wide total, so the real total is simply unknown there, not zero; see note for what's actually known.
jobs_matching
Optional
_type: company_no_matching_jobs rows only. Always 0 -- the reason the row exists. Kept as an explicit field so you can check one field name across both job rows (absent) and these rows (0) when scanning for "any jobs".
reason
Optional
_type: company_no_matching_jobs rows only. Why there are zero matching jobs: no_jobs_found (site read fine, genuinely no open positions), filtered_out_by_filters (jobs exist but none matched your keywords/postedWithinDays), or no_location_match_on_site (jobs exist but none are in your requested location).
note
Optional
_type: company_no_matching_jobs rows only. Short human-readable explanation of reason, naming the company and (when relevant) your active filters.
filters
Optional
_type: company_no_matching_jobs rows only. The keywords/location/postedWithinDays/titleFilter active on THIS run, echoed so you can tell "no match against these filters" apart from a genuinely job-less company without re-reading the run's input: {keywords, location, posted_within_days, title_filter}.
scraped_at
Optional
_type: company_no_matching_jobs rows only. ISO8601 timestamp of when job-feed observed this company's zero-match result (not when the sub-actor itself extracted it).
