Workable Job Search MCP Server

Workable Jobs Scraper & API - All Companies, Remote, 24 Hours

Workable jobs scraper and API: search every company hiring on Workable (apply.workable.com and jobs.workable.com) at once and get each job's title, company, every place, remote or not, job type, posting date, company website and links, newest first. Turn on full details for the description, requirements and benefits, the pay when the company shows it, the department and the company's own job code.

Use it as a Workable API for every company's jobs, an apply.workable.com scraper, a jobs.workable.com scraper or a daily Workable job feed from the Apify Console, the Apify API, Python, JavaScript, n8n, Make or an AI agent, and export JSON, CSV or Excel. It is the site:apply.workable.com Google dork done at the source, with the fields a search result cannot give you. Available as an Apify Actor. $0.0025 per job, $0.0035 with full details, $0.01 per run start, no login, no API key.

What does it do?

• Searches every company on Workable at once. A word, a place, a work type or a job type reads Workable's own feed of every job its customers publish on its job board, which Workable updates every hour. Workable's job board counted 170,374 listings on 2026-11-02, one per job and place; the first three quarters of the feed alone held 79,763 open jobs at 4,792 companies in 181 countries, among them Rentokil Initial, Amartha, SAP Fioneer, Teleperformance Spain, Biffa, Foodics, Blueground and Hugging Face.
• One keyword is enough. Type a job title and press Start.
• Newest first, read at run time. Every job comes from Workable during your run, not from a stored copy of ours.
• Filters that take what people type. Germany, Texas or TX, EMEA, Remote, contract, internship, last 24 hours, a jobs.workable.com search link, or a pasted Google dork such as site:apply.workable.com intext:engineer.
• Any company by name. Type Hugging Face, the account in its address or its career page link, and that career page is read live.
• Checks job links. apply.workable.com and jobs.workable.com links come back open, closed or not found.

How is it different from other Workable scrapers?

• You do not need to know the companies. Half of the Workable-only actors below read Workable's job board search, which ranks by relevance, 20 jobs a page, with no date order; the others need each company's name or link, or read a stored copy. This actor reads Workable's own feed of every company's jobs and sorts the matches by date. On 329 hiring career pages checked on 2026-11-02, the feed held 7,691 of their 7,761 open jobs (99%). On Apify the whole feed arrived on 2026-11-03, with jobs listed two hours before the run; from our laptop Workable cut 15 of 16 downloads about a minute in: see Limitations.
• One row per job. Workable lists a job once per place. Here a job's places come together in one row: 14,534 of 79,763 jobs (18%) had two places or more on 2026-11-02.
• The real posting date. A search by word carries the time Workable listed the job, which fell on the job page's own day for 95% of 6,000 jobs checked (98.5% of those published since September). With full details on, and for the companies you name, a search's rows carry the day the job page itself gives Google (its datePosted, 10 of 10 pages checked).
• Details from each job's own answer: the full description, requirements and benefits, pay as numbers with its period when the company shows it, remote, hybrid or onsite, the department and the company's job code. Workable's job board shows no pay.
• You pay per saved job, plus one run start, and a spending limit stops the run cleanly with the jobs it paid for.

When should I use it?

• Job seekers and job boards: remote and international jobs at 4,792 companies in 181 countries, on a daily schedule. In the first three quarters of the feed on 2026-11-02, the US held 43% of the jobs, the UK and Greece 6% each, Canada 4%.
• Recruiters and staffing agencies: see who is hiring for a role across every company on Workable.
• Sales and market research: every job found by a word names the hiring company and its website, a ready list of companies that are hiring now.
• Data teams and AI agents: one clean, stable feed of Workable jobs through the API or MCP.

How much does it cost to scrape Workable jobs?

Pay per event: $2.50 per 1,000 jobs, or $3.50 per 1,000 with full details, plus $0.01 per run at the default 1 GB of memory. Apify's free $5 monthly credit covers about 1,900 jobs in runs of 100. Platform usage is included, and there is no monthly fee.

The Store page header, "from $2.00 / 1,000 jobs", is the GOLD price: Apify's Store shows the lowest discount-tier price. FREE and BRONZE (Apify's Free and Starter plans) pay the base price, $0.0025 a job and $0.0035 with details; SILVER (Scale) pays 10% less, $0.00225 and $0.00315; GOLD (Business) and above pay 20% less, $0.002 and $0.0028. The run start, $0.01 at the default 1 GB, is a flat fee on every plan, and the header never shows it.

Why a run start: every search by word reads Workable's whole feed, about 140 MB compressed, whatever it finds, and the run start covers that fixed work. Number of jobs (maxJobs) and the run start set the most a run can cost, and a run stops at its spending limit with a message saying so.

How does the price compare with other Workable scrapers?

Two groups, from the 3,522 job listings the Apify Store returned for 37 Workable and multi-ATS terms on 2026-11-02 (1,353 of them mention Workable), each ordered by users in the last 30 days, with free-plan prices from each actor's pricing record.

Every actor made for Workable alone with 2 or more users:

Every actor for many platforms that reads Workable, with 40 or more users:

Not counted: kaix/linkedin-jobs-scraper (174 users) reads LinkedIn and names Workable only as an apply link, doggo/uk-jobs-board-scraper (40) checks apply links against ATS names, and parsebird/hiring-cafe-scraper (39) reads hiring.cafe.

Where others are cheaper: shahidirfan and schnellscrapers charge $0.001 a job plus a $0.005 start through Workable's job board search, api-co (1 user) $0.00075 on the free plan and $0.0012 to $0.0015 on paid plans the same way, practical_ophthalmologist_iuq $0.0018 plus a $0.001 start and deadwood_data_solutions $0.002 plus $0.01 a query for the career pages you name, jobspipe $0.00196 in its many-board search, and openclawai $0.0005 when you know the companies. Where others do more: the multi-platform actors cover many ATS in one run, and jobo.world, factpipe and jtpalms put the description in every row.

Which inputs does it take?

Nothing is required. Type a job title and press Start, or press Start with the form empty.

Example input:

{
  "searchTerms": ["software engineer"],
  "location": ["Germany", "United States"],
  "postedWithin": "30 days",
  "maxJobs": 100
}

What does the output look like?

One row per job, newest first. These five rows (ranked 1, 9, 22, 45 and 69 of 100) are from a run on Apify on 2026-11-03, build 1.0.1, searching every company for software engineer in 157 seconds at the default 1 GB of memory: 1,177 jobs at 488 companies, the 100 newest saved, the newest listed two hours before the run.

One row as JSON, from the same run:

{
  "title": "Senior Software Engineer - AI Oversight",
  "company": "OpenMined",
  "location": "Colombia",
  "otherLocations": ["Brazil", "Argentina", "Uruguay"],
  "workType": "remote",
  "employmentType": "contract",
  "department": null,
  "salary": null,
  "postedAt": "2026-11-02T23:26:54+00:00",
  "jobUrl": "https://apply.workable.com/j/5DAF8993BF",
  "applyUrl": "https://apply.workable.com/j/5DAF8993BF/apply",
  "status": "open",
  "platform": "workable",
  "companySlug": "openmined",
  "companyWebsite": "http://openmined.org",
  "jobId": "5DAF8993BF",
  "checkedAt": "2026-11-03T09:11:49+00:00"
}

Full details add description, descriptionHtml, reference, salaryMin, salaryMax, salaryCurrency, salaryPeriod, city, region, country and countryCode, and fill department, salary and the work type from each job's own answer. One real row with pay, from a local run on 2026-11-02: Project Engineer - Construction at Path Construction in Chicago, onsite, "salary": "USD 65,000 - 80,000 per year", "salaryMin": 65000, "salaryMax": 80000, "salaryCurrency": "USD", "salaryPeriod": "year". A checked link that is closed or not found comes back with its status and link, and the other fields empty. A run also writes an OUTPUT record: whether Workable's feed or the career pages were read, how many jobs matched and were saved, and any notes on how your input was read.

How fast is it?

About 2.5 minutes at the default 1 GB of memory, for a search by word and for the empty form alike (build 1.0.1, 2026-11-03). Software engineer across every company took 157 seconds, almost all of it reading Workable's whole feed, which arrived in full (run 6uFK1C6wG1NyXXCyO). The empty form took 155 seconds for the 300 biggest career pages, all of which answered (run 7XF3iMyDlPMqHCzpz). Each saved 100 jobs: $0.26 with the run start.

How the time goes: a search by word, place, work type or job type is one download of Workable's feed, about 140 MB compressed and 860 MB of XML, read once for every word. On Apify the whole feed arrived. From our laptop on 2026-11-02 and 2026-11-03, Workable ended that answer about 63 seconds after the request on 15 of 16 downloads; when it is cut, the run uses the part that came and 120 seconds of career pages. An empty search reads the career pages of the 300 biggest companies on our list, 2 a second, about 150 seconds, and reads the feed instead when Workable refuses them before any answers. Full details, a hybrid or onsite search and the companies you name ask Workable 2 calls a second, at most 1,000 calls a run.

Common questions

How does it search every company without a list?

The actor searches every company through Workable's own feed of "all job information for all our customers published on Jobs by Workable", updated every hour, as its help center says. A search by word, place, work type or job type reads that feed, so no list of companies is needed: on 329 hiring career pages checked on 2026-11-02, the feed held 99% of their open jobs. The companies you name, the empty form and job links read each company's own career page; for those the actor carries a list of 2,196 Workable accounts built from Common Crawl, the Wayback Machine and Workable's own feed, each of which answered on its career page on 2026-11-02 or 2026-11-03. A company that is not on the list is still read when you paste its link.

Which posting date does it use?

The actor's posting date (postedAt) in a search by word is the time Workable listed the job on its job board. Measured on 6,000 jobs read both in the feed and on their career page on 2026-11-02, that was the job page's own day for 95% of them (98.5% of those published since September); 3.5% were listed more than a week after they were published, mostly old jobs listed late. The companies you name, and every search row with full details, carry the day the job page gives Google as its datePosted, with no time of day. A checked jobs.workable.com link carries the time Workable's job board gives it.

Why does a hybrid or onsite search return fewer jobs?

Workable's feed and career pages mark only remote jobs. On Workable's own job board on 2026-11-02, 16.6% of listings were hybrid and 54.6% onsite, but neither word is in the feed. So a hybrid or onsite search asks each matching job's own Workable answer, newest first, 2 calls a second, for up to 90 seconds, and the run's message says how many it checked. A remote search needs no check.

Does it show pay?

The actor shows pay only with full details on, and only when the company shows pay on the job: Workable's feed and its job board carry none. Pay comes as Workable states it, with the currency and the period (year, month, week, day or hour).

Why did a run say Workable refused it?

Workable's Cloudflare refuses an address that calls apply.workable.com too often, for up to a day (measured on 2026-11-02). The actor paces its calls at 2 a second and makes at most 1,000 in a run, and once Workable refuses it, it stops asking at once and delivers the jobs it already read. An empty search refused before any career page answers reads Workable's feed instead, which comes from www.workable.com. Apify's addresses are shared with other users, so run again later, or search by a word.

Does it need a login or an API key?

The actor needs no login and no API key. Workable's official API needs each employer's own token and reads only that employer's jobs. This actor reads what Workable publishes for anyone: its jobs feed, each company's career page and its job board. You need only an Apify account.

Does it work on Apify's free plan?

The actor works on Apify's free plan: the free $5 monthly credit pays for about 1,900 jobs in runs of 100, at $0.26 a run with the run start.

How often should I run it?

Running the actor once a day keeps a Workable job feed fresh. A search by word reads Workable's feed, which Workable updates every hour, so runs less than an hour apart read the same feed. Give the run a word, place, work type or job type so it reads every company (the empty form reads only the 300 biggest career pages) and set Posted in the last (postedWithin) to 24 hours. Set the run up under Schedules in Apify Console, and match rows on jobId to skip jobs you already have.

How do I get every job of one company?

To get every job of one Workable company, type its name in Only these companies (companies), such as Hugging Face, the account in its address (huggingface) or its Workable link, and set Number of jobs (maxJobs) to 0 for every open job. Its career page is read live, with the department and the company's job code. A company that is not on our list of 2,196 accounts is tried by its name in lower case with hyphens; paste its link if that misses.

How do I check a list of job links?

To check Workable job links, paste them into Job links to check (jobUrls), one per line, instead of searching: apply.workable.com, old company.workable.com and jobs.workable.com links all work. Each link comes back as one row, open with the job's fields, or closed or not found with its link, so 10 links come to $0.035 with the run start. On apply.workable.com a closed job and one that never existed answer the same, so both come back closed.

How do I call it from the API, Python, JavaScript, n8n or an AI agent?

Calling the actor from the Apify API, Python, JavaScript, n8n, Make or an AI agent takes your Apify token (Apify Console, Settings, API & Integrations) and the input as JSON, with the field names above.

API. One call runs the actor and returns its rows. Apify stops that wait after 300 seconds: a search by word takes about 2.5 minutes, inside it, but full details ask Workable up to 1,000 times at 2 calls a second, so for a details run use a client below, which waits until the run ends.

curl -X POST "https://api.apify.com/v2/acts/anshumanatrey~workable-job-search/run-sync-get-dataset-items" \
  -H "Authorization: Bearer <YOUR_APIFY_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{"searchTerms": ["software engineer"]}'

Python (pip install apify-client):

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("anshumanatrey/workable-job-search").call(run_input={"searchTerms": ["software engineer"]})
for row in client.dataset(run.default_dataset_id).iterate_items()  # apify-client 3+; on 2.x use run["defaultDatasetId"]:
    print(row["title"], row["company"], row["location"], row["postedAt"], row["jobUrl"])

JavaScript (npm install apify-client):

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });
const run = await client.actor('anshumanatrey/workable-job-search').call({ searchTerms: ['software engineer'] });
const { items } = await client.dataset(run.defaultDatasetId).listItems();
for (const row of items) console.log(row.title, row.company, row.location, row.postedAt, row.jobUrl);

n8n and Make. In n8n, add the Apify node, pick Run an Actor and get dataset, choose this actor and paste the input JSON: the rows come out as n8n items. In Make, the Apify app's Run an Actor and Get Dataset Items modules do the same.

AI agents (MCP). Add https://mcp.apify.com?tools=anshumanatrey/workable-job-search to your MCP client and sign in with Apify, or send Authorization: Bearer <YOUR_APIFY_TOKEN>. The agent gets this actor as a tool and sees every input field by its JSON key.

Am I charged for closed jobs?

A closed job is charged only when you paste its link to check: a search returns open jobs only, and a pasted link that turns out closed or not found is one Job row, because it was checked and answered.

Is it legal to use?

The actor reads public job postings. You are responsible for how you use the data, including each site's terms and the privacy laws that apply to you.

Limitations

• An empty search does not read every company: it reads the career pages of the 300 biggest companies on our list, about 150 seconds. Type a word, a place, a work type or a job type to search every company.
• Workable's feed holds only the jobs a company publishes on Workable's job board: 99% of the open jobs on the career pages checked.
• Workable ends the feed's download about a minute after it starts. On Apify on 2026-11-03 the whole feed arrived; from our laptop on 2026-11-02 and 2026-11-03 it cut 15 of 16 downloads, the furthest reaching the listings of 2026-10-22. If a run's download is cut, the run keeps every job listed before the cut, reads 120 seconds of the career pages with the most matching jobs for the newer ones, names the cut date in its message and marks OUTPUT partial. Other companies' newest jobs are not in such a run.
• Hybrid and onsite are checked job by job, so those searches return fewer jobs (see above).
• Full details ask Workable once or twice per job, at most 1,000 times a run: about 650 jobs found by a word get details, and the rest come without them, at the Job price.
• Pay comes only with full details, when the company shows it; jobs.workable.com links carry no pay.
• About a fifth of jobs state no job type (19.9% on 2026-11-02), so a job type filter leaves them out.
• The department and the company's job code come only from career pages or with full details: Workable's feed has neither.
• The Workable account (companySlug) is known for 54.8% of the jobs found by a word; every job found by a word has the company's name and website.
• On apply.workable.com a closed job and one that never existed answer the same, so both read as closed; jobs.workable.com links tell them apart.
• A company that is not on our list is tried by its name in lower case with hyphens, which can miss: paste its link instead.
• Workable can refuse an address for up to a day (see above).
• The list of 2,196 accounts is partial: each answered on its career page on 2026-11-02 or 2026-11-03, but Workable banned our laptop before most candidate accounts were checked. A search by word, place, work type or job type does not depend on it.
• Titles and descriptions come in the company's own language.

Who maintains it?

Built and maintained by Anshuman Atrey (@AnshumanAtrey), who publishes a portfolio of data and OSINT actors on Apify Store. For a feature request or a bug, message LinkedIn (fastest), open an issue on the GitHub repo, or write through atrey.dev.

Sibling actors

For the people behind a company: LinkedIn Profile Finder turns an email into a public LinkedIn profile.

Documentation

• Apify Store: https://apify.com/anshumanatrey/workable-job-search
• GitHub repo: https://github.com/AnshumanAtrey/workable-job-search (MIT license, see LICENSE)
• The engine behind it: https://github.com/AnshumanAtrey/ats-jobs-engine

Last updated

2026-11-03 (version 1.0)