Job-Room Scraper

Stop re-checking Job-Room by hand: get only the new jobs for your searches every morning, in a sheet, Slack or your own job board. Search Job-Room (Switzerland, all language regions) by keyword or location and get every matching job as clean JSON with the salary split into min / max / interval; schedule it and each run marks which jobs are new since your last run.

How "new since last run" works

Don't know which companies to watch? The Greenhouse Jobs Scraper searches 4,800 companies by job title, with no company names needed, and reports exactly which matching roles are new or closed each run.

The Actor remembers which jobs each search returned last time. A job it has not seen before is marked isNewSinceLastRun: true. If the whole search fits within Max jobs, the comparison is exact and jobs that disappeared come back as type: "closed" rows. If the search is bigger than Max jobs, a job only counts as new when it was also posted after your previous run, so an old job that merely moved into view is never reported as new.

Filters (title, location text, salary, remote) never change the comparison, so changing a filter never makes jobs look new.

What you get

Field
title, company, location, canton	as listed, with postcode and canton code (company is null when the employer posts anonymously)
latitude, longitude	coordinates of the workplace
workloadMinPercent, workloadMaxPercent, employmentType	the workload the Swiss way (80–100%), and full-time / part-time read from it
contractDuration, startDate, endDate	permanent or fixed-term, immediately or a date, and the end of a fixed-term job
reportingObligation, reportingObligationEndDate	true for jobs under the Swiss mandatory job reporting rule (Stellenmeldepflicht): occupations with high unemployment, shown first to job seekers registered with an RAV — no other job board has this flag
occupationCode, experience, qualification, languages	the RAV occupation code, the experience asked for (3+ years…), skill level and the languages the employer lists
description, descriptionLanguage	the full job text (German, French, Italian or English, as posted)
numberOfJobs, postedAt, validUntil, updatedAt	openings, publication start and end, last change
externalUrl, applyUrl	the employer's own posting and its application form
isNewSinceLastRun	true / false, or null on the first run (unknown, not "old")
jobUrl	link to the job on Job-Room
search	which of your searches returned it

Contact persons, their phone numbers and e-mail addresses are never returned. Job-Room has no salary field, so there are no salary columns.

Input

Put one search per line in Searches (e.g. Pflege) and optionally a Location (e.g. ZH). Location takes a canton code (ZH, GE, VD…) or a town or postcode, which is matched in the job's place. Many searches in one run are cheaper than one run each.

A search the site refuses comes back as an error row and the run carries on. When the job board itself is down or failing (HTTP 5xx or no answer), the error row says so and carries sourceDown: true, so you can tell it from a problem in this Actor; nothing is charged for what could not be read.

Get only the new jobs every morning (3 steps)

1. Open this Actor in Apify Console, fill in your searches, set What to return to Only jobs that are new since the last run, and click Save as a new task.
2. On the task, add a Schedule, for example daily at 07:00.
3. Open the task's Integrations tab and connect Slack, email, Google Sheets (or Zapier / Make / a webhook). Each morning you get only the new jobs.

You pay only for the new jobs plus $0.01 per run, so a quiet day costs about a cent. When you export by hand, pick the Jobs view: 12 tidy columns, title first, instead of every raw field.

Pricing

$0.70 per 1,000 jobs, plus $0.01 per run. Closed-job rows and errors are free. With Only jobs that are new since the last run, a daily schedule only pays for the new ones.

For AI agents

For a fast answer call with {"searches": ["Pflege"], "maxJobsPerSearch": 20, "includeDescription": false}. It takes a few seconds. The defaults (up to 200 jobs with full descriptions) can take about 2 minutes, which is longer than many agent tool timeouts. Each row has type: job, closed or error. isNewSinceLastRun: null means "no previous run to compare with".

Use it from Claude Code (or another coding agent)

Save this as ~/.claude/skills/job-room/SKILL.md (other agents that read SKILL.md files: their skills folder), put your Apify API token in the APIFY_TOKEN environment variable, and ask the agent in plain words, for example "Any new Pflege jobs on Job-Room since yesterday?".

---
name: job-room
description: Job listings from Job-Room by keyword, location and filters, and only the jobs that are new since the last run. Runs the Apify Actor northbell/job-room-scraper with the user's APIFY_TOKEN.
---
Run: curl -s -X POST "https://api.apify.com/v2/acts/northbell~job-room-scraper/run-sync-get-dataset-items?timeout=300" -H "Authorization: Bearer $APIFY_TOKEN" -H "Content-Type: application/json" -d '{"searches":["Pflege"],"maxJobsPerSearch":50,"output":"new"}'
"output": "new" returns only jobs not seen in the previous run (the first run returns all). Use "all" for a one-off list.
Rows have type job, closed or error. Never print the token. Tell the user the price on the Actor's Store page before a large run.

Prefer MCP? Add https://mcp.apify.com?tools=northbell/job-room-scraper as a remote MCP server (in Claude Code: claude mcp add --transport http apify "https://mcp.apify.com?tools=northbell/job-room-scraper") and sign in to Apify when asked; the Actor then shows up as a tool.

Same filters on every site (10/2026): experience level (read from the title: intern → executive, returned as a seniority column) and employment type (the site's own words normalised into full-time / part-time / contract / temporary / internship, returned as employmentTypes). They only narrow what is returned, with no extra requests and no extra cost.

Zero rows is never ambiguous. The run's status message says why, per search, e.g. "zzz": no results on the site, results found but all removed by your filters or nothing new since the last run, and the OUTPUT record in the run's key-value store has the same as JSON: perSearch[].status is ok, no-results, all-filtered, nothing-new or error, with listed, read, returned, new and closed. Set summaryRows: true to get it as free type: "summary" rows instead.

Same core columns on every northbell job scraper (15 job sites + 10 company-career-page scrapers), so you can stack them into one table: jobId, title, company, location, countryCode (ISO, job sites), remote (true / false / null), employmentType (the site's own words), postedAt, jobUrl, salaryMin, salaryMax, salaryCurrency, salaryInterval, isNewSinceLastRun. Unknown is null, never a guess.

More job scrapers from northbell

Job sites: SEEK · Instahyre · Reed · InfoJobs · XING · Dice · HelloWork · Computrabajo · Foundit · jobs.ch · Arbeitsagentur · Job-Room · Platsbanken · eFinancialCareers · MyCareersFuture · EURES · USAJOBS · Job Bank · Internshala · Gupy · CutShort · iimjobs & hirist · Naukrigulf · Jobindex · karriere.at · Arbeitnow · Wellfound · Remote OK · We Work Remotely · Just Join IT & No Fluff Jobs · OnlineJobs.ph.

Company career pages with exact new/closed tracking: Greenhouse · Lever · Workday · AI companies.