SEEK Jobs Scraper (Australia & New Zealand)

💼 What is SEEK Jobs Scraper?

SEEK Jobs Scraper extracts job ads from SEEK in Australia and New Zealand: title, company, salary (hidden ones inferred), full description, number of applicants, screening questions, contact e-mails and phones, closing date and company profile, for any keyword, location or SEEK search URL.

Type a keyword (nurse) and a location (All Sydney NSW), or paste any SEEK search URL, click Start, and download the jobs in JSON, CSV or Excel. No login, nothing to set up, and it is fast: about 100 job ads with full details in a minute. Every job of a search, even beyond the 550 that SEEK shows.

📋 What data can you extract from SEEK?

One item per job ad, 91 fields:

Category	What you get
🏷️ Job	title, SEEK's standard role name, link, apply link, job id — Management Accountant
💰 Salary	salary as shown, min / max, per hour / year, currency, super included — and the range of a hidden salary, inferred from SEEK's own salary filter — $65 - $75 per hour + super
👥 Applicants	number of applicants, application volume, share with a resume / a cover letter — 3 applicants, Low application volume
📝 Ad	full description (text and HTML), summary, selling points, screening questions, video, cover image
☎️ Contacts	e-mail addresses and phone numbers given in the ad
🧭 Signals	the level of the job (entry, junior, senior, lead, manager, executive), whether the ad offers or refuses visa sponsorship, and a 0-100 score of its good signs — worked out from the ad by rules, included in the price
🤝 Recruiter	for agency ads, the recruiter's SEEK profile: name, rating, number of reviews, phone, agency, website, location, specialisations, placements — 4.96 / 5 from 32 reviews
🏢 Company	company, advertiser, verified badge, registration date, company profile (industry, size, website, description, head office), rating, number of reviews, salary rating, perks, awards, open jobs, and optionally the latest employee reviews
🗂️ Category	SEEK classification and sub-classification, work type (full time, part time, contract, casual), on-site / hybrid / remote
📍 Location	location, suburb, area, state, postcode, country
🕒 Dates	listing date, closing date, expired or not

Every field, with an example, is listed in the Output section below.

Fields marked detail in the Output tab (description, applicants, screening questions, contacts, recruiter, closing date, company profile) are filled when Extract details is ticked (off by default: it is charged per job). Leave it off for a results-only scrape that is several times faster.

✅ Why use SEEK Jobs Scraper?

• 🔓 Every job of a search: SEEK shows at most 550 jobs per search; this Actor splits a bigger search on its own (by category, work type, state…) and returns all of them — 1,500 « nurse » jobs from all Australia in 1 to 2 minutes, results only.
• 💰 Hidden salaries inferred: about half of SEEK's ads hide the salary; the Actor finds its range with SEEK's own salary filter (salaryIsInferred: true), charged only when a range is found (tick Infer hidden salaries; off by default).
• 👥 Applicants and screening questions: how many people applied, and what the employer asks when you apply.
• 🚀 Fast: about 100 job ads with full details in a minute, 1,200 in under 10 minutes, 1,500 results-only jobs in 1 to 2 minutes.
• 🇦🇺 🇳🇿 Australia and New Zealand in one Actor, with SEEK's own filters: classification, sub-classification, work type, work arrangement, salary, listing date, advertiser, company, radius.
• 🔔 Monitoring built in: tick Only new jobs, schedule the Actor, and each run returns (and charges) only the jobs it has never delivered before.
• 🔌 API, scheduling, integrations (Make, Zapier, n8n, Google Sheets…) and JSON / CSV / Excel export via the Apify platform.

🚀 How to scrape SEEK

1. Create a free Apify account.
2. Open SEEK Jobs Scraper, type Search keywords (e.g. nurse) and a Location (e.g. All Sydney NSW), and pick the Country.
3. Or paste your own SEEK URLs into Start URLs: any search results page (the filters set on SEEK are kept) or single job ads.
4. Set Max jobs (100 by default, 0 = every job of the search) — and Max jobs per search when you run several searches — then click Start.
5. Download the dataset in JSON, CSV, Excel or via API.

💰 How much does it cost to scrape SEEK?

This Actor uses pay per event pricing: $0.18 per 1,000 job ads with full details, or $0.17 per 1,000 from the results only (Extract details unticked, the default; a job pasted by its URL or id is always read in full) — plus $0.01 per run start. Higher Apify plans pay less, down to $0.15 and $0.14 per 1,000. Company open-job counts are included, and so are the signals (seniority, visaSponsorship, jobScore). The text of employee reviews (Max reviews per company, off by default) is charged per review delivered, $0.79 per 1,000 reviews (down to $0.76 on higher plans), once per company per run (event company-review, "Company review"): the first job of a company that carries its reviews pays them, the next jobs of that company carry the same reviews at no cost. Add market statistics per search (unticked by default) is charged per search delivered, with at least 10 jobs, $0.015 per search ($14.99 per 1,000 searches, down to $14.96 on higher plans; event market-stats, "Market statistics"). A hidden salary inferred is charged per job whose range is found, $2.79 per 1,000 on top of the job price, down to $2.76 on higher plans (event inferred-salary, "Hidden salary inferred"; a salary shown, or one SEEK's bands cannot place, is never charged; Infer hidden salaries is unticked by default: tick it to have them found). The post-filters are applied by the Actor on every job it reads: each job a filter checks is charged, kept or not (event filter-check when the results card answers it — dates, title / excluded / included words, salary; detail-filter-check when the job ad must be read — e-mail, phone, verified advertiser, applicants, included words with Extract details).

Example: 10,000 job ads with full details ≈ $1.81; a daily monitor of 300 new jobs with full details ≈ $0.06. Platform usage (compute, proxy) is included in the price.

⚙️ Input

{
    "query": "nurse",
    "location": "All Sydney NSW",
    "maxItems": 200
}

Several searches, a cap per search, SEEK filters, only recent jobs, only the ones not delivered before:

{
    "searchQueries": ["registered nurse", "midwife"],
    "locations": ["All Sydney NSW", "All Melbourne VIC"],
    "maxItemsPerQuery": 100,
    "workType": ["full-time"],
    "workArrangement": ["hybrid", "remote"],
    "postedAfter": "7 days",
    "onlyNew": true,
    "stateKey": "nurses-au"
}

Or with your own URLs and job ids:

{
    "startUrls": [
        { "url": "https://www.seek.com.au/nurse-jobs/in-All-Sydney-NSW?worktype=242" },
        { "url": "https://www.seek.co.nz/jobs?keywords=developer&where=Auckland" }
    ],
    "jobIds": ["94713920"],
    "maxItems": 500
}

Field	Notes
query, searchQueries	Job title, skill or company, exactly as typed on SEEK (nurse -aged, nurse OR midwife, "registered nurse" work); searchQueries adds more keywords (one search each). Ignored when startUrls is set.
location, locations	Suburb, postcode, city, region or state as SEEK names it (All Sydney NSW, 2000, New South Wales NSW, Auckland); locations adds more places. Every keyword is searched in every location (max 500 searches per run). Empty = the whole country.
country	AU (seek.com.au, default) or NZ (seek.co.nz).
radius	km around a precise location (suburb, postcode); 0 = SEEK's default (50 km).
startUrls	SEEK search results pages (filters kept, pagination automatic) or single job ads, from seek.com.au or seek.co.nz.
jobIds	SEEK job ids (94713920) to extract directly — AU and NZ ids mixed: each is read on its own site, whatever country.
maxItems	Stop after this many jobs for the whole run (0 = unlimited).
maxItemsPerQuery	Cap for EACH search (keyword × location, or search URL). 0 = no per-search cap.
classification	SEEK categories (Healthcare & Medical, Information & Communication Technology…).
subClassification	SEEK sub-categories by name (Web Development & Production), by Classification > Name when several share it (Healthcare & Medical > Management), or by id (6302).
workType	full-time, part-time, contract, casual.
workArrangement	on-site, hybrid, remote.
salaryType	Unit of the salary filter: annual, monthly, hourly.
salaryMin, salaryMax	SEEK's salary filter; it also matches the ads that hide their salary.
dateRange	SEEK's date filter, in calendar days of Sydney time (on seek.co.nz too): 1 = today since midnight, 3, 7, 14, 31 = since midnight that many days ago, or any. For the last 24 hours, use postedAfter.
sortBy	date (newest first, default) or relevance.
advertiserIds	Only the jobs of these SEEK advertisers (the advertiserId of a result).
companyName	Only the jobs of this company (Ramsay Health Care).
extractDetails	Open each job ad for the full description, applicants, questions, contacts and company (off by default; charged per job read, see the pricing).
inferHiddenSalary	Infer the range of a hidden salary from SEEK's salary filter (off by default; charged per salary found, see the pricing).
maxReviewsPerCompany	Latest employee reviews added to each job (field companyReviews), read once per company; 0 = none. Charged per review delivered, once per company per run (see the pricing).
includeMarketStats	Adds the MARKET_STATS record to the run's key-value store: one sheet of figures per search (off by default; charged per search delivered, see the pricing and Market stats per search below).
postedAfter, postedBefore	Listing date range: 2026-09-01, or a period before now (7 days, 2 weeks, 1 month, 24 hours).
titleKeywords, includeKeywords, excludeKeywords	Keep the jobs whose title contains one of the words; whose title, company, teaser, selling points (or, with Extract details, description) contain one; drop the ones that contain one (case and accents ignored).
requireEmail, requirePhone, onlyWithSalary, verifiedOnly	Keep only the jobs with a contact e-mail, a phone number, a salary (shown or inferred), a verified advertiser.
minApplicants, maxApplicants	Keep the jobs with at least / at most this many applicants (at most 10 = little competition).
onlyNew, stateKey, resetState	Monitoring: only the jobs never delivered under this memory key; resetState forgets the memory.
Advanced	proxyConfiguration (Apify proxy by default, included in the price; the residential proxy is not available), maxConcurrency, maxRequestsPerMinute, minRequestIntervalMs (0 by default: no gap between two requests; both count the results pages and job ads, not the salary bands nor the company counts and reviews), maxRequestRetries, debugLog.

📦 Output

One real item from a test run (long texts shortened; the recruiter's name and contacts replaced):

{
    "id": "94824689",
    "url": "https://www.seek.com.au/job/94824689",
    "applyUrl": "https://www.seek.com.au/job/94824689/apply",
    "title": "Accounts & Payroll Officer - Civil Construction",
    "normalisedRoleTitle": "Payroll and Accounts Officer",
    "roleId": "payroll-and-accounts-officer",
    "teaser": "Join a Growing Melbourne Civil Construction business in a hands-on Accounts role covering payroll, b…",
    "bulletPoints": [
        "Growing & Established Melbourne Civil Construction business",
        "Own this Weekly-Payroll, Accounts, Reconciliations & Monthly Reporting tasks"
    ],
    "abstract": "Join a Growing Melbourne Civil Construction business in a hands-on Accounts role covering payroll, b…",
    "description": "The Company\nOur client's a well-established subcontractor servicing some of Australia's largest Tier…",
    "descriptionHtml": "<h2><strong>The Company</strong></h2><p>Our client's a well-established subcontractor servicing some…",
    "companyName": "Cappro Pacific",
    "advertiserId": "46218953",
    "advertiserName": "Cappro Pacific",
    "advertiserIsVerified": true,
    "advertiserRegistrationDate": "2020-12-03T02:06:44.057Z",
    "isPrivateAdvertiser": false,
    "companyProfileId": null,
    "companyProfileUrl": null,
    "companyLogoUrl": "https://bx-branding-gateway.cloud.seek.com.au/e5113d33-7dfa-461d-a4ef-3196e3270871.1/jdpLogo",
    "companyIndustry": null,
    "companySize": null,
    "companyWebsite": null,
    "companyDescription": null,
    "companyRating": null,
    "companyReviewCount": null,
    "companyPerks": [],
    "companyReviews": [],
    "employerId": "465831",
    "companyOpenJobs": 19,
    "companyPrimaryLocation": null,
    "companySpecialities": null,
    "companySalaryRating": null,
    "companyAwards": [],
    "classificationId": "1200",
    "classification": "Accounting",
    "subClassificationId": "6140",
    "subClassification": "Accounts Officers/Clerks",
    "workType": "Full time",
    "workArrangements": [
        "On-site"
    ],
    "isRemote": false,
    "isHybrid": false,
    "location": "Melbourne VIC",
    "suburb": null,
    "area": "Melbourne VIC",
    "state": "Victoria",
    "postalCode": null,
    "countryCode": "AU",
    "salaryLabel": "$100,000–$120,000 per year plus Super",
    "salaryMin": 100000,
    "salaryMax": 120000,
    "salaryPeriod": "year",
    "salaryCurrency": "AUD",
    "salaryIncludesSuper": true,
    "salaryIsInferred": false,
    "publishedAt": "2026-09-23T04:47:49.000Z",
    "listingDateDisplay": "6m ago",
    "expiresAt": "2026-10-23T12:59:59.999Z",
    "isExpired": false,
    "applicantCount": 1,
    "applicantVolumeLabel": "Low application volume",
    "applicantsWithResumePercentage": 100,
    "applicantsWithCoverLetterPercentage": 100,
    "screeningQuestions": [],
    "contactEmails": [
        "jane.doe@acme-recruit.com.au"
    ],
    "contactPhones": [
        "0412 345 678"
    ],
    "seniority": null,
    "visaSponsorship": null,
    "jobScore": 100,
    "hasRoleRequirements": false,
    "applicationIsExternal": false,
    "isFeatured": false,
    "isSponsored": false,
    "tags": [
        "Be an early applicant"
    ],
    "badges": [
        "On-site"
    ],
    "videoUrl": null,
    "coverImageUrl": "https://image-service-cdn.seek.com.au/fafb8e81ada107b6cfcb07d366e4626315f95472/00e3f857d81b293723f9026c526521e4285c2beb",
    "recruiterProfileUrl": "https://au.seek.com/recruiters/people/jane-doe",
    "recruiterName": "Jane Doe",
    "recruiterRating": 4.86,
    "recruiterReviewCount": 87,
    "recruiterPhone": "0412 345 678",
    "recruiterAgency": "Cappro Pacific",
    "recruiterWebsite": "https://www.cappropacific.com.au/",
    "recruiterLocation": "Sydney, NSW, 2213, Australia",
    "recruiterSpecialisations": [
        "Engineering",
        "Sales",
        "Mining, Resources & Energy"
    ],
    "recruiterPlacements": 85,
    "searchQuery": "accountant",
    "searchLocation": null,
    "searchUrl": "https://www.seek.com.au/jobs?keywords=accountant&sortmode=ListedDate&page=1",
    "scrapedAt": "2026-09-23T04:54:11.718Z"
}

You can download the dataset in various formats such as JSON, HTML, CSV or Excel.

All 91 fields

Fields	What you get
id, url, applyUrl, title, normalisedRoleTitle, roleId	Job: 94820549, job and apply links, Management Accountant
teaser, bulletPoints, abstract, description, descriptionHtml, videoUrl, coverImageUrl	Ad: summary, selling points, full text and HTML, video and cover image
salaryLabel, salaryMin, salaryMax, salaryPeriod, salaryCurrency, salaryIncludesSuper, salaryIsInferred	Salary: $65 - $75 per hour + super → 65, 75, hour, AUD, true; salaryIsInferred: true when the range comes from SEEK's filter
applicantCount, applicantVolumeLabel, applicantsWithResumePercentage, applicantsWithCoverLetterPercentage, screeningQuestions, hasRoleRequirements, applicationIsExternal	Applicants: 3, Low application volume, 100, 33, the employer's questions
contactEmails, contactPhones	Contacts given in the ad
seniority, visaSponsorship, jobScore	Signals worked out from the ad by rules (see the FAQ): senior, false (the ad refuses sponsorship) or true (it offers it), 80
recruiterProfileUrl, recruiterName, recruiterRating, recruiterReviewCount, recruiterPhone, recruiterAgency, recruiterWebsite, recruiterLocation, recruiterSpecialisations, recruiterPlacements	Recruiter of an agency ad: profile link, 4.96 / 5 from 32 reviews, agency, Accounting & Finance, placements
companyName, advertiserId, advertiserName, advertiserIsVerified, advertiserRegistrationDate, isPrivateAdvertiser, employerId, companyOpenJobs	Advertiser: Perigon Group, verified, registered in 2010, 41 open jobs
companyProfileId, companyProfileUrl, companyLogoUrl, companyIndustry, companySize, companyWebsite, companyDescription, companyPrimaryLocation, companySpecialities	Company profile on SEEK
companyRating, companyReviewCount, companySalaryRating, companyPerks, companyAwards, companyReviews	Company reputation: 3.4 / 5 from 183 reviews, perks, awards, the latest reviews (title, rating, pros, cons, job title, date)
classificationId, classification, subClassificationId, subClassification, workType, workArrangements, isRemote, isHybrid	Category: Accounting, Management Accounting & Budgeting, Contract/Temp, On-site
location, suburb, area, state, postalCode, countryCode	Location: Brisbane QLD, Brisbane, Brisbane QLD (the SEEK region, with its state code), Queensland, 4000, AU
publishedAt, listingDateDisplay, expiresAt, isExpired	Dates: listing date (what the date filters read), closing date
isFeatured, isSponsored, tags, badges	SEEK's featured / sponsored flags, tags and badges
searchQuery, searchLocation, searchUrl, scrapedAt	The search that found the job, ISO timestamp

💡 Tips

How to get more results

Set maxItems to 0: a search above SEEK's 550 jobs is split on its own, and every job is returned once. A broad keyword (or none, with a category) is fine.

How to reduce costs

The price is per job, so the levers are maxItems, maxItemsPerQuery, the SEEK filters, the post-filters (a filtered-out job is not saved, but each job a filter drops is charged a check — a job it keeps pays its own price only: prefer the SEEK filters, applied by SEEK itself) and onlyNew for recurring runs (you never pay twice for the same job).

Market stats per search (optional)

Tick Add market statistics per search (includeMarketStats) and, once the run has delivered its jobs, one record named MARKET_STATS is added to the run's key-value store (Output tab) — never to the dataset. It holds one sheet per search, worked out from the jobs of this run only:

• the jobs delivered, and the jobs SEEK announced for the search;
• the share of jobs with a salary shown, and with a salary inferred;
• salary median and quartiles for each pay period (year, month, day, hour), in the most frequent currency, on the middle of each shown range;
• the share of remote and hybrid jobs, the work types, the 10 employers who advertise most, the top classifications and sub-classifications;
• the levels (seniority), the share of ads that offer visa sponsorship, the median number of applicants (jobs read with their ad), the share of jobs listed in the last 7 days.

A search needs at least 10 jobs delivered; a job pasted by its URL or id belongs to no search. With Only new jobs, only the jobs delivered in this run count. Charged per search delivered, see the pricing.

Several searches in one run

Fill searchQueries and / or locations: the Actor runs one search per keyword × location (3 keywords × 4 cities = 12 searches, up to 500 per run). The single query and location fields still work and are added to the lists. A job found by several searches is saved — and charged — once. Set maxItemsPerQuery to give every search its own cap: without it the first searches can use up the whole maxItems budget. You can also paste several search URLs into startUrls: each one is a search of its own, with the same cap.

Monitoring: only the new jobs

Tick Only new jobs (onlyNew) and schedule the Actor. The first run returns everything; each later run skips the jobs already delivered: they are not saved, not charged, and their ad is not even opened. The memory lives in a named key-value store of your account (seek-jobs-scraper-seen, up to 150,000 jobs per key) and is only updated with jobs that really reached the dataset, so a failed run never hides anything. Give each schedule its own stateKey (two schedules sharing a key would hide each other's jobs), and tick resetState once to start over. Searches are then sorted by newest, and a search stops once it meets 100 jobs in a row you already have, or fewer when Max jobs or Max jobs per search is smaller (the sponsored jobs SEEK pins at the top are not counted). A search the Actor splits (above SEEK's limit) also skips the jobs listed before the previous run that its maxItems left out: a capped run does not hand you the older rest of the search next time. A job your filters drop is checked (and charged the filter check) once: the next runs with the same filters skip it for free; change a filter, or tick resetState, to check them again.

Filter by publication date

postedAfter and postedBefore take a date (2026-09-01, the whole day is included: Sydney time for a job of seek.com.au, Auckland time for seek.co.nz) or a period before now (7 days, 2 weeks, 1 month; via the API also 24 hours or a full ISO date-time). The filter reads publishedAt; a job without a listing date is dropped as soon as a date bound is set. Filtered-out jobs are not saved and do not count in maxItems (each job the filter checks is charged, see the pricing); the run summary tells how many were filtered. With postedAfter, searches are sorted by newest and stop at the first page that is entirely too old.

🔌 Integrations and API

Call the Actor via the Apify API, the JavaScript or Python clients, or connect it with integrations and webhooks (Make, Zapier, n8n, Google Sheets, Slack, Airtable…). The dataset can be fetched as JSON or CSV from any tool.

🤖 Use with AI agents (MCP)

AI agents (Claude, ChatGPT, Cursor…) can find and run this Actor through the Apify MCP server, billed to their Apify account like any run. It returns one item per SEEK job ad (Australia or New Zealand). Actor id: nice_dev/seek-jobs-scraper; MCP server with this Actor only: https://mcp.apify.com/?tools=fetch-actor-details,nice_dev/seek-jobs-scraper.

Smallest input, for a cheap first call:

{
    "query": "nurse",
    "location": "All Sydney NSW",
    "maxItems": 10
}

Key output fields: url, title, companyName, location, salaryLabel, workType, publishedAt, applicantCount.

Cost: $0.18 per 1,000 job ads with full details ($0.17 from the results only) plus $0.01 per run start, down to $0.15 and $0.14 per 1,000 on higher Apify plans; the inferred hidden salaries ($2.79 per 1,000, down to $2.76) cost extra, see the pricing section above. Cap each call with maxItems and, through the API, with the run option maxTotalChargeUsd.

❓ FAQ

Is it legal to scrape SEEK?

The Actor only reads what SEEK shows publicly to any anonymous visitor. It logs in to nothing and solves no captcha. Results can contain personal data (a recruiter's name, e-mail or phone number written in an ad, anonymous employee reviews), which is protected by GDPR and the Australian Privacy Act: do not store it without a legitimate reason. You are responsible for using the data in compliance with SEEK's Terms of Use and applicable law. This Actor is not affiliated with SEEK.

Does it need a login or a proxy?

No login. The proxy is included in the price: leave the default setting (the residential proxy is not available). A request the site turns away is retried at once on a new proxy session (without a proxy, after a pause of 5 seconds, doubled at each retry up to 150 seconds). If SEEK still turns away 10 job ads in a row after all their retries, the run stops and its status says so, keeping the jobs saved until then: run it again later.

How is a hidden salary inferred?

When an advertiser hides the salary, SEEK still filters on the range they typed. The Actor asks SEEK the same search once per salary band ($30k, $40k, … $350k per year) and keeps the bands the job comes back in: salaryMin / salaryMax are those bands' bounds, per year, with salaryIsInferred: true. The range is as precise as SEEK's bands (often $20k-$50k wide). Only a range found is charged (event inferred-salary), once its job is saved.

How are seniority, visa sponsorship and the job score worked out?

By fixed rules, not by a model, from what the ad says. seniority is read in the job title (else in SEEK's standard role name), on whole words: graduate, intern, trainee… are entry; junior; senior; lead, principal, team leader; manager, head of, supervisor; director, chief, CEO, general manager… are executive. A title with several takes the highest (« Senior Account Manager » is manager); one with none gives null. visaSponsorship is false when the ad refuses sponsorship or asks for working rights or citizenship (« no sponsorship », « must have full working rights », « Australian citizens only »), true when it offers it (« visa sponsorship available », « will sponsor », « 482 visa »), null when it says nothing; it is read in the description, or in the summary and selling points of a job whose ad was not opened. jobScore is the share of good signs among those known, from 0 to 100: a salary shown, a job listed at most 3 days before it was read, a contact e-mail or phone given (known once the ad is opened), fewer than 20 applicants, a verified advertiser.

Is the data safe to open in Excel or to show on a web page?

Titles and descriptions are the advertisers' own words, copied as they are. A text can begin with -, +, = or @: Excel and Google Sheets may read such a cell of a CSV file as a formula or as a number. The Actor leaves the text as it is, so that the JSON and the API give the real value: when you open a CSV, import these columns as text. descriptionHtml is the ad's HTML as SEEK serves it, not sanitized: on a web page, escape every field like any text written by a stranger. Every link field (url, companyWebsite, videoUrl, recruiterWebsite…) is an http(s) URL or null: never javascript: or data:, never a quote, never a user name before the host (https://www.seek.com.au@other.example is dropped).

Known limitations

• SEEK filters that are not in the input can still be used by pasting a filtered search URL into startUrls.
• applicantCount is only shown by SEEK for ads you apply to on SEEK, once a few people have applied; it is null otherwise.
• onlyNew remembers job ids, not their content: a job whose salary changed is not returned again.
• Two runs sharing the same stateKey at the same time may both return the same new job.
• A split search returns what SEEK gives for each part: SEEK adds loose keyword matches to a broad search that it leaves out of the narrower ones (measured: 1.7 % of « nurse » healthcare jobs, e.g. a GP job), so a few of those are missing.
• Sorted by date with a cap, a search gives its newest jobs up to about 490. Above that the Actor splits it (by category, work type…) and each part gives its own newest ones: the jobs returned are not the newest of the whole search, and do not come out in date order.
• A search the Actor splits (more than about 490 jobs wanted from a search of more than 500), added later to a stateKey already used, only returns the jobs listed since that key's last run: give it its own stateKey.

A run the platform stops without warning (out of memory, run timeout)

• Resurrect it: it goes on from where it stood at most a minute before the stop. What it had read since is read again, and the jobs already saved are skipped: none is delivered or charged twice, and maxItems still counts them.
• The jobs found in the last 2 seconds before the stop may be missing from the dataset: they are not charged.
• With onlyNew, the memory is saved after each write (at most once every 2 seconds): resurrect the stopped run and the jobs it had saved meanwhile join the memory; leave it stopped for good, and the next run may return the jobs of its last 2 seconds once more.

Something doesn't work?

The last line of the log counts the jobs saved, filtered out and no longer on SEEK (removed while the run was reading them), and the requests that failed after every retry. Those requests and the removed jobs are listed, with the reason, in the FAILED_REQUESTS record of the run's key-value store. A run that saved nothing and had failed requests fails, and its last message gives the cause. A run that saved some jobs fails too when at least as many requests failed for good as were read (SEEK refusing the job ads after the first ones, for instance): a green run with a short dataset would hide the outage. One failed request among many is only a warning.

If SEEK changes its pages, you are told instead of paying for blank rows. A results page that counts jobs but gives none the Actor can read is an error (listed in FAILED_REQUESTS), never a quiet "No listings found". If the first 20 jobs read all lack their title, listing date, classification or — with extractDetails — description, the run saves nothing more, stops and fails, and its last message names the missing field: at most those first jobs are charged.

🛟 Support

Open an issue in the Issues tab with a link to your run: the run log and the FAILED_REQUESTS record of the key-value store show exactly which searches or jobs failed and why.