JOIN Jobs Search Scraper
Pricing
from $2.99 / 1,000 join job records
JOIN Jobs Search Scraper
Extract public job postings from Join.com company career pages with title, locations, description HTML, public dates, and source URLs.
Pricing
from $2.99 / 1,000 join job records
Rating
0.0
(0)
Developer
Jobs API
Maintained by CommunityActor stats
0
Bookmarked
3
Total users
2
Monthly active users
16 days ago
Last modified
Categories
Features
Share
This Actor collects source-grounded job and employer information published on public JOIN company career pages and public JOIN job pages. It reads the page's public __NEXT_DATA__ payload; it does not use an authenticated account, private API, proxy, cookies, browser fingerprint changes, or CAPTCHA/security-challenge workarounds.
Input
startUrls accepts one to twenty HTTPS URLs on join.com (or www.join.com) in one of these forms:
https://join.com/companies/{company}https://join.com/companies/{company}/{job}
If omitted, the Actor uses the public JOIN company page https://join.com/companies/join. An explicitly supplied invalid URL is rejected; it is not silently replaced with the default.
includeDetails defaults to true and fetches the public job detail pages needed for full descriptions and job attributes. The dataset contract requires a successful detail response and at least 21 distinct meaningful source fields, so includeDetails: false intentionally emits no listing-only rows. maxItems (1–100, default 20) limits emitted jobs across all input URLs. maxPages (1–10, default 3) bounds listing pages per company URL. requestTimeoutSeconds (5–120, default 30) and maxResponseBytes (100,000–5,000,000, default 5,000,000) bound each ordinary public request.
Example:
{"startUrls": [{ "url": "https://join.com/companies/join" }],"includeDetails": true,"maxItems": 20,"maxPages": 3}
Dataset
Each dataset item is a complete, detail-verified job record—not a listing fallback or a run diagnostic. A row is emitted only after a successful 2xx detail response and when it contains at least 21 distinct meaningful populated source-field paths. Identifiers, URLs, aliases, repeated description renderings, raw source snapshots, and run metadata do not count toward this threshold. Where the source publishes them, records retain the job title and status, dates, employer identity and description, industry and employee range, offices, mission, benefits and values, structured location and remote/hybrid terms, work schedule and language, employment and contract types, category, salary range/currency/frequency, and readable source-authored description sections. Optional facts are omitted when JOIN does not publish them; fields are not filled with guesses.
Diagnostics and request receipts are stored in KVS (RUN_DIAGNOSTICS and REQUEST_RECEIPTS), never as dataset rows. OUTPUT_SUMMARY.status is the platform process status (SUCCEEDED or FAILED); resultStatus separately reports COMPLETE, LIMITED, SKIPPED, DEFERRED, NO_DATA, or FAILED. A handled source barrier (HTTP 401/403/429/451 or visible source-authored denial/challenge) discards every buffered row and returns SUCCEEDED / SKIPPED. A generic transport failure is DEFERRED, not proof of a block; if complete rows were already stored, they are preserved as SUCCEEDED / LIMITED. Likewise, a later dataset-write or finalization error cannot relabel previously stored rows as failed. A fatal zero-row run calls the Actor failure lifecycle and reports FAILED / FAILED; a clean run without eligible job rows reports SUCCEEDED / NO_DATA. No diagnostic or placeholder item is added to the dataset. The Actor does not retry blocked requests or change identities, cookies, proxies, or fingerprints. For ordinary navigation only, it follows at most one site-authored redirect when the destination remains on an allowed JOIN /companies/{same-company} page or job path; both hops are retained as receipts. This handles JOIN's public canonical job links and is never used after a block or challenge.
You can download the dataset in various formats such as JSON, HTML, CSV, or Excel.
Local checks
From this directory, run npm test, npm run check, npm run lint, and apify validate-schema. Dataset validation is read-only. Do not run an Actor just to validate an existing store; when an Actor run is explicitly in scope, choose a fresh actor-relative storage directory, preserve all existing stores, and use --resurrect (never --purge):
$env:APIFY_LOCAL_STORAGE_DIR = "storage-join-validation-unique"if (Test-Path -LiteralPath $env:APIFY_LOCAL_STORAGE_DIR) { throw "Choose a new storage directory." }apify run --resurrect --input-file .actor/input.jsonnpm run validate:dataset -- storage-join-validation-unique
The dataset validator reads generated output only; it does not create or edit dataset records.