Research & Academic Jobs Scraper — 10 Sources

One call, ten sources. Fans out to 10 actors covering academia, UN/NGO and policy hiring, merges and dedupes into one dataset.

Two sources need your own Anthropic or Mistral key. devex and ub_doctoral extract listings using an AI provider (Claude/Anthropic or Mistral). Pass anthropicApiKey or mistralApiKey in the input to include them — either key is enough; Anthropic is preferred when both are set. Without either key, those two sources are skipped — you are not charged for them — the other 8 sources run normally, and the output includes a warning row for each skipped source (see BYOK behavior below).

What research jobs data does this scraper extract?

Each result is one flat JSON record per job posting:

Field	Type	Meaning
source	string	Which child board the record came from, e.g. "reliefweb"
id	string	Stable source-side identifier ("" when the source has none)
title	string	Job title as posted
company	string	Hiring company / organisation
location	string	Location / duty station (may include remote hints)
url	string	Direct link to the posting
postedAt	string	Posting date where the source provides it, else ""
deadline	string	Application deadline where the source provides it, else "" — populated by euraxess, jobs_ac_uk, un_careers, reliefweb, devex, math_ku_phd and ub_doctoral; linkedin, impactpool and ikerbasque don't expose a deadline, so those always return ""
snippet	string	Short description excerpt
salary	string	Salary text where the source provides it, else "" — only jobs_ac_uk populates this today; the other 9 sources always return ""

Warning rows (see below) additionally carry a warning field and every other field is null instead of "".

How the bundle works

This is a bundle Actor: one endpoint that fans out to the individual job-source Actors listed below, runs them concurrently, maps every record onto one flat schema and dedupes by URL across boards. You can restrict the run to a subset with the sources input. Each child source is charged its own pay-per-event pricing on top of this bundle's — that is the cost of one-call breadth.

BYOK behavior (Devex & UB doctoral)

devex and ub_doctoral read listings that need extra rendering and use an AI provider (Anthropic or Mistral) to extract structured data from them, so they require anthropicApiKey or mistralApiKey — either one is enough; when both are set, Anthropic is used and mistralApiKey/mistralModel are ignored for these two sources. If you don't pass either:

• Both sources are removed from the run before any child actor is started — you are not billed their $0.05 actor-start fee.
• One warning row per skipped source is pushed to the dataset: {"source": "devex", "warning": "skipped: requires anthropicApiKey or mistralApiKey", ...other fields null...}.
• The run's status message (visible in the Apify Console and via the Runs API) also states which sources were skipped and why.
• This applies even if you explicitly list devex or ub_doctoral in sources — without a key they still can't produce rows, so they're still skipped and still get a warning row.
• The other 8 sources are completely unaffected.

Your Anthropic or Mistral usage for these two sources is billed separately, directly by that provider — it is not part of this Actor's Apify pricing.

How to scrape research jobs with this Actor

1. Click Try for free / Run — no login to the target site, no cookies, no proxies to configure.
2. Adjust the input (keyword, filters, maxItems) or keep the defaults.
3. Run it and export the dataset as JSON, CSV or Excel, or read it over the API.

Run it from your own code:

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("nomad-agent/researcher-bundle").call(run_input={"maxItems": 50})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["title"], "—", item["company"], item["url"])

Or a single HTTP call that runs the Actor and returns items in one response:

curl -X POST \
  "https://api.apify.com/v2/acts/nomad-agent~researcher-bundle/run-sync-get-dataset-items?token=<YOUR_APIFY_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{"maxItems": 50}'

Input

Field	Type	Default	Notes
sources	array	["linkedin", "euraxess", "jobs_ac_uk", "un_careers", "reliefweb", "impactpool", "ikerbasque", "devex", "math_ku_phd", "ub_doctoral"]	Which boards to include. Leave empty to run the full default set. Each enabled source costs its own $0.05 actor-start plus per-result fees.
keyword	string	""	Optional free-text filter forwarded to children that support it (others ignore it).
anthropicApiKey	string (secret)	""	Needed (or mistralApiKey, below — either is enough) only for devex and ub_doctoral. Without either, those two are skipped with a warning row — see BYOK behavior. Billed separately by Anthropic. Preferred over mistralApiKey when both are set.
mistralApiKey	string (secret)	""	Alternative to anthropicApiKey for devex and ub_doctoral — pass this instead if you'd rather bring a Mistral key. Billed separately by Mistral.
mistralModel	string	"mistral-small-latest"	Mistral model forwarded to devex/ub_doctoral when they run on mistralApiKey (no anthropicApiKey supplied).
maxItemsPerSource	integer	36	Cap on items fetched from EACH child board before merge.
maxItems	integer	360	Hard cap on the merged, deduped output. Default is sources × maxItemsPerSource (the zero-config ceiling). Set 0 for no cap.
cacheTtlSeconds	integer	1800	How long to reuse results already fetched from a source instead of re-fetching. 0 = always fetch fresh.
concurrency	integer	6	How many child boards to run in parallel. (Advanced)
runTimeoutSecs	integer	120	How long to wait for each source before giving up on it. (Advanced)
apifyToken	string (secret)	""	Leave empty — injected automatically on the Apify platform. Only set for local runs outside the platform. (Advanced)
actorOwner	string	""	Which Apify account's child actors to call. Leave empty to use this bundle's published sources. (Advanced)

Output example

A normal record:

{
  "source": "reliefweb",
  "id": "4218889",
  "title": "Protection Officer",
  "company": "Danish Refugee Council",
  "location": "Ukraine",
  "url": "https://reliefweb.int/job/4218889/protection-officer",
  "postedAt": "2026-06-28",
  "deadline": "",
  "snippet": "The Protection Officer supports...",
  "salary": ""
}

A source populating deadline (euraxess, jobs_ac_uk, un_careers, reliefweb, devex, math_ku_phd, ub_doctoral) or salary (jobs_ac_uk only) fills in that field instead of "". A warning row, pushed when a BYOK source is skipped:

{
  "source": "devex",
  "id": null,
  "title": null,
  "company": null,
  "location": null,
  "url": null,
  "postedAt": null,
  "deadline": null,
  "snippet": null,
  "salary": null,
  "warning": "skipped: requires anthropicApiKey or mistralApiKey"
}

Pricing

Pay per event: $0.05 per Actor start and $0.004 per job returned — plus each enabled child source's own pay-per-event pricing (also $0.05 per start + $0.004 per result, charged by that child actor directly).

Zero-config run estimate (defaults, no anthropicApiKey/mistralApiKey): 8 runnable sources × up to 36 items = up to ~288 merged items, roughly $2.75 all-in ($0.05 + 8×$0.05 in child starts + up to 288 items × $0.008 combined per-result fee). Pass an anthropicApiKey (or mistralApiKey) to also run devex and ub_doctoral: 10 sources × up to 36 items = up to ~360 items, roughly $3.43 all-in. Real runs usually cost less — not every board returns the full cap, and cross-board duplicates are billed once by whichever child returned them first but not double-billed by the bundle.

Skipped BYOK sources (neither key provided) cost nothing — no child actor-start, no per-result fees, and the warning row itself is not billed. Your own Anthropic or Mistral API usage (when you supply a key) is billed separately, directly by that provider.

Use cases

• Academic job boards and PhD/postdoc alert bots
• NGO/policy career services
• Research-mobility analytics
• University career offices

FAQ

Is it legal to scrape research jobs? This Actor reads only publicly available job postings — data any visitor can see without logging in. No personal data behind authentication is touched. Review the target site's terms and your local regulations for your specific use case.

Do I need an account on the target site? No. Postings are fetched from public pages/APIs — no login, cookies or session tokens.

How fresh is the data? Every run fetches live listings. Results are cached for cacheTtlSeconds (default 30 min, set 0 to always hit the source live).

How many jobs can I get? maxItems caps the run (set 0 for no cap). Most sources paginate from newest to oldest.

Why don't I see Devex or UB doctoral jobs? You didn't pass anthropicApiKey or mistralApiKey — see BYOK behavior above. Add either key to include those two sources.

Something broken or missing? Open an issue on the Actor's Issues tab — it is monitored and reliability fixes ship fast.

Integrations

Export the dataset as JSON, CSV or Excel, or read it straight from the Apify API. Works out of the box with Make, Zapier and n8n via their Apify integrations, can be called synchronously with run-sync-get-dataset-items from any backend, and is usable by AI agents through the Apify MCP server.

Related Actors

• LinkedIn Jobs Scraper — No Login, No Cookies
• EURAXESS Jobs Scraper — EU Research Positions
• jobs.ac.uk Scraper — UK Academic & Research Jobs
• UN Careers Job Scraper — United Nations Jobs
• ReliefWeb Jobs Scraper — Humanitarian & NGO
• Impactpool Jobs Scraper — UN & NGO Careers
• Ikerbasque Jobs Scraper — Basque Research Roles
• Devex Jobs Scraper — International Development
• University of Copenhagen PhD Jobs Scraper (KU)
• Universitat de Barcelona PhD Vacancy Scraper