LIVE SCRAPER

Jobs

Thirteen Polish job sites in one search: Pracuj.pl, Just Join IT, No Fluff Jobs, OLX Praca, LinkedIn Jobs, Indeed, praca.pl, theprotocol.it, RocketJobs, Bulldogjob, Solid.Jobs, the labour offices’ CBOP database and the civil-service recruitments at KPRM. Every site is queried in parallel and answers for itself in sources[] with a status (ok, no_results, error, timeout or stopped), so one slow or blocked board never fails the call. Offers come back in one shape (salary with its period, contract and gross flag; contracts as b2b, uop, uz, uod, internship), interleaved best-first, and the same job found on several boards appears once with its other copies in also_on. A filter a site cannot apply is applied after the fetch or named in ignored_filters, never dropped silently. Six sites are fetched directly and free (Solid.Jobs, No Fluff Jobs, Bulldogjob and CBOP in about a second, praca.pl in 2–4 s, KPRM in 10–17 s with a keyword), OLX Praca and LinkedIn go direct first with an unblocker fallback, and five render through a paid unblocker (Pracuj.pl, theprotocol.it, Indeed, Just Join IT, RocketJobs; Indeed falls back to Google Jobs through SerpAPI). Each source has about 30 s (62 s for the three heavy pages), results are cached 10 minutes, and a site that blocks is left alone for 15 minutes and reported as stopped.

3 tools ~1–30 s cache: 10 min PLN 13 job sites: 8 fetched directly (2 with an unblocker fallback), 5 through an unblocker

Data from Pracuj.pl, Just Join IT, No Fluff Jobs, OLX, LinkedIn, Indeed, praca.pl, theprotocol.it, RocketJobs, Bulldogjob, Solid.Jobs, CBOP (oferty.praca.gov.pl) and KPRM (nabory.kprm.gov.pl); Indeed results may come from Google Jobs; rights belong to their operators. Unofficial, not affiliated. Legal notice

Thirty seconds, any client.

claude mcp add --transport http jobs https://mcp.wondel.ai/jobs/mcp

One command in your terminal — the server is available in every Claude Code session.

This server works with Claude, Claude Code, Claude Desktop, Claude Cowork, Cursor, VS Code, Windsurf — and any other MCP client. Stdio-only clients connect through npx mcp-remote.

By connecting you accept the legal notice: personal use, data and rights stay with the source sites.

3 tools, fully typed.

Parameters below are transcribed from the server’s own zod schemas — what you see is what the tool validates.

jobs_search Job offers across the 13 sites in one call, each queried in parallel and merged into one list: the best offer of each source first, duplicates folded into also_on. Pass keyword and/or location; narrow with sources or category instead of asking every site. sources[] reports each site’s status, total, backend and the filters it could not apply.
Param Type Description
keyword string Job title, skill or phrase, e.g. "python developer", "księgowa", "kierowca C+E". Needed unless location is set.
location string A Polish city or region as people write it, e.g. "Warszawa", "Kraków". Needed unless keyword is set.
remote boolean true = remote work only. Some sources cannot filter it (see ignored_filters).
seniority enum Experience level. internjuniormidseniorlead
contract enum[] Contract types: b2b, uop (umowa o pracę), uz (zlecenie), uod (o dzieło), internship. b2buopuzuodinternship
salary_min number Minimum monthly gross salary in PLN, above 0, e.g. 15000.
with_salary boolean true = only offers that publish a salary.
sources string[] Source ids to query (default: all): pracuj, justjoin, nofluffjobs, olx, linkedin, indeed, pracapl, theprotocol, rocketjobs, bulldogjob, solidjobs, cbop, kprm. Mutually exclusive with category.
category enum Query every source of a kind: it (tech boards), general, public (public sector). Mutually exclusive with sources. itgeneralpublic
page integer
= 1
Result page per source, 1–20.
per_source integer
= 10
Offers wanted from each source, 1–25.
sort enum
= relevance
relevance: sources interleaved, each in its own ranking; newest: by posted_at, newest first. relevancenewest
no_cache boolean
= false
Bypass the 10-minute result cache and the 15-minute pause on blocked sources.
jobs_offer One offer’s full text: the description as plain text, requirements and benefits when the site separates them, the apply link and the same fields as a search offer. Pass the url from jobs_search, or source with id. Solid.Jobs has no per-offer endpoint, so its offers open only shortly after the jobs_search that returned them; an OLX Praca offer may be refused from the server’s network (then open the url yourself).
Param Type Description
url string The offer’s url from jobs_search. Use this, or source + id.
source string Source id (pracuj, justjoin, nofluffjobs, olx, linkedin, indeed, pracapl, theprotocol, rocketjobs, bulldogjob, solidjobs, cbop, kprm), with id.
id string The offer’s id from jobs_search, with source.
no_cache boolean
= false
Bypass the result cache.
jobs_sources The sources jobs_search can query: id (for sources), site, kind (free = fast and unmetered; paid = through an unblocker, slower), categories (it, general, public), which filters each one supports, typical latency and the fetch route. Static and free: call it to choose sources without guessing.
Param Type Description

Prompts that just work.

Find senior Python jobs in Warszawa or remote with a published salary of at least 25 000 PLN a month, and tell me which ones appear on more than one board.

jobs_search

Which public-sector offers for an accountant are open in Kraków right now, and what does the KPRM one pay?

jobs_search + jobs_offer

I only want the boards that answer in a second. Which ones are those, and do any of them filter by contract type?

jobs_sources

What comes back

jobs_search · live response, captured 2026-11-07 from production, abridged

{
  "fetchedAt": "2026-11-07T11:46:15.161Z",
  "query": { "keyword": "python", "location": "Warszawa", "page": 1, "per_source": 2, "sources": ["pracuj", "justjoin", "nofluffjobs", "olx", "linkedin", "indeed", "pracapl", "theprotocol", "rocketjobs", "bulldogjob", "solidjobs", "cbop", "kprm"], "sort": "relevance" },
  "sources": [
    { "source": "pracuj", "name": "Pracuj.pl", "status": "ok", "total": 714, "count": 2, "cached": false, "backend": "firecrawl+brightdata", "ignored_filters": [] },
    { "source": "justjoin", "name": "Just Join IT", "status": "ok", "total": 1125, "count": 2, "cached": false, "backend": "firecrawl+brightdata", "ignored_filters": [] },
    { "source": "nofluffjobs", "name": "No Fluff Jobs", "status": "ok", "total": 336, "count": 2, "cached": false, "backend": "direct", "ignored_filters": [] },
    { "source": "indeed", "name": "Indeed (pl)", "status": "ok", "total": 1221, "count": 2, "cached": false, "backend": "firecrawl+brightdata", "ignored_filters": ["seniority", "contract"] },
    { "source": "cbop", "name": "CBOP (oferty.praca.gov.pl)", "status": "ok", "total": 1, "count": 1, "cached": false, "backend": "direct", "ignored_filters": [], "note": "location matched \"Warszawa, m.st. Warszawa, mazowieckie\" (exact locality, no radius)" },
    …
  ],
  "count": 24,
  "offers": [
    {
      "source": "pracuj",
      "id": "ai-engineer-python-llms-warszawa,oferta,1005176980",
      "url": "https://www.pracuj.pl/praca/ai-engineer-python-llms-warszawa,oferta,1005176980",
      "title": "AI Engineer (Python, LLMs)",
      "company": "Addepto",
      "locations": ["Warszawa"],
      "work_mode": "office",
      "salary": [{ "from": 12600, "to": 19320, "currency": "PLN", "period": "month", "contract": null, "gross": null }],
      "contract": ["uz", "b2b"],
      "seniority": "mid",
      "posted_at": "2026-11-07T11:31:58.273Z",
      "expires_at": "2026-11-22T22:59:59Z",
      "tags": [],
      "snippet": "Your responsibilities, Support the development of scalable AI systems, including LLMs and GenAI, …"
    },
    {
      "source": "justjoin",
      "id": "dcg-python-developer-warszawa-python-e1a068e3",
      "url": "https://justjoin.it/job-offer/dcg-python-developer-warszawa-python-e1a068e3",
      "title": "Python Developer",
      "company": "DCG",
      "locations": ["Warszawa"],
      "work_mode": "hybrid",
      "salary": [{ "from": 140, "to": 150, "currency": "PLN", "period": "hour", "contract": "b2b", "gross": false }],
      "contract": ["b2b"],
      "seniority": "mid",
      "posted_at": "2026-11-03T09:05:05.114075Z",
      "expires_at": "2026-11-27T22:59:59.999999Z",
      "tags": ["Python", "Mikroserwisy", "Linux", "Cloud"],
      "snippet": null,
      "also_on": [{ "source": "nofluffjobs", "url": "https://nofluffjobs.com/pl/job/python-developer-dcg-warszawa-3" }]
    },
    …
  ],
  "note": "Filters some sources could not apply: seniority (Indeed (pl)); contract (Indeed (pl))."
}

jobs_search returns one envelope: fetchedAt says when, query echoes what you asked with the defaults filled in, and sources[] stands in for a single source field, one entry per site with its status (ok, no_results, error, timeout or stopped), the site’s own total, how many offers it contributed, cached, the fetch chain in backend and the filters it could not apply. jobs_offer carries fetchedAt and backend on the offer itself; jobs_sources is static. Read salaries as the site shows them: period is hour, day, month or year (null when the site does not say), gross is false for B2B net rates and null when the site does not say, and contract is null when a range is not tied to one contract. A mistake in your input (an unknown source id, sources and category together, an offer url no site owns) gets a guided error that names the fix and ends with a link to this page. A blocked site is not a mistake of yours: it shows as stopped in sources[] for 15 minutes while the other sites still answer. Compact JSON in one text block, exactly how agents like it.

Add Jobs
to your agent.

mcp.wondel.ai/jobs/mcp

Authless · Streamable HTTP at /mcp · Unofficial, for personal use under the legal notice