Skip to main content
GET
Search Google Jobs
Use mode=serp for public job listings from employers and job boards. The default mode=rpc searches Google’s own Careers openings. In SERP mode, description contains the full text Google includes for that listing, including the part hidden behind “Show more,” with its paragraph and line breaks. A few listings carry only a short text in Google itself; for those, use apply_links to open the original posting. job_highlights contains Google’s structured qualifications, responsibilities, and benefits when present; it can be null. There is currently no separate public Google Jobs detail endpoint. Switching to mode=rpc does not expand an aggregator listing; it searches a different source. Google’s own per-job detail response can contain the same truncated preview. To extract text from the original posting, you can make a separate Web Scraping API request to an apply_links URL. That request uses Web Scraping pricing and depends on the job board’s accessibility. Each Jobs search costs 20 credits per request, not per returned job. In SERP mode, the API attempts to recover missing application links from Google’s detail response for the same job ID, within the existing request deadline. This recovery adds no customer credit charge. url is the Google listing URL; use apply_links or apply_options[].link for application destinations. If Google does not supply a recoverable destination, those arrays can still be empty; the API does not invent a link or substitute a different listing. Google Jobs only lists jobs near where it believes the searcher is. In SERP mode the search is located in the gl (or country) country, so q=camarero&gl=es returns jobs in Spain without adding “España” to the query. A place in q or location still narrows the search, for example marinero Cádiz. To pick a different location, send Google’s uule value; it replaces the country default.

More results (pagination)

Each SERP page returns about 10 jobs. When Google has more, pagination.next holds an opaque token such as 2.EswB…. Send it back as next_page_token, with the same q, gl and hl, to get the next page; pagination.current is the page number. When pagination.next is null, there are no more results. Each page is one search (20 credits).

Filters

Google ignores most job_type and ltype values now; date_posted still works. filters lists Google’s own filter options for the search, each with a name (for example “Tiempo parcial” or “Última semana”), a refined q and a uds token. To apply one, repeat the request with that q and uds.

Job fields

Besides title, company_name, location, via, description and the apply links, each job carries the facts Google shows on its card: posted_at, salary (as Google writes it, for example “De 30 mil € a 46 mil € al año”), schedule_type (also in job_type), and detected_extensions with the same facts plus work_from_home and qualifications when present. job_id is Google’s stable listing ID, useful for removing duplicates across pages and searches. A successful empty jobs array means Google explicitly reported no matching jobs for that location. Blocked pages and failed fetches return an error. Requests use a bounded retrieval window of up to 45 seconds; allow at least 60 seconds for your client’s timeout, including network overhead.

Authorizations

x-api-key
string
header
required

Query Parameters

q
string
required

Job title / keywords / combined query.

location
string | null

City / state / region — concatenated with q before sending to Google. The search is also located in the gl/country country.

gl
string
default:us

Country code (ISO 3166 alpha-2). In mode=serp the search is also located in this country (Google Jobs only lists jobs near where it believes the searcher is), unless you send uule. A place named in q or location still narrows it.

country
string | null

Alias for gl; when present overrides it.

hl
string
default:en

Language code.

language
string | null

Alias for hl; when present overrides it.

domain
string
default:google.com

Google domain for locale-specific results (google.com, google.co.uk, google.co.in, …).

job_type
enum<string> | null

Employment type — translated into a chips filter. Google often ignores it now; use a filters entry (q + uds) from a previous response instead.

Available options:
FULLTIME,
PARTTIME,
CONTRACTOR,
INTERN
date_posted
enum<string> | null

Posted-date window — translated into a chips filter.

Available options:
today,
3days,
week,
month
ltype
enum<string> | null

Work arrangement — maps onto Google's remote/hybrid/onsite chips.

Available options:
remote,
hybrid,
onsite,
work_from_home
chips
string | null

Raw Google chip-filter string (comma-separated). Merged with any structured filters (job_type, date_posted, ltype).

uds
string | null

Opaque Google filter token harvested from a prior Jobs search URL.

uule
string | null

Google's UULE-encoded location (e.g. w+CAIQICIFU3BhaW4 for Spain). Overrides the default gl/country location in mode=serp.

lrad
string | null

Search radius around the location (Google accepts a distance in miles).

next_page_token
string | null

Opaque token from the previous response's pagination.next (SERP mode: <page>.<cursor>); send it with the same q, gl and hl for the next page.

mode
enum<string>
default:rpc

Data source. rpc (default, ~300 ms) replays Google's own r06xKb batchexecute RPC on the Google Careers portal — clean JSON, 20 roles per page, scope = Google's internal openings. serp uses the public Jobs search vertical (udm=8, SERP-embedded, 3rd-party aggregator) and uses HTTP retrieval with browser fallback inside a 45-second budget. Descriptions contain the full text Google supplies; apply_links point to the original listings.

Available options:
rpc,
serp

Response

Successful Response

Response for GET /api/v1/jobs/search.

pagination
Pagination · object
required

Pagination metadata.

jobs
JobResult · object[]
jobs_results
JobResult · object[]
filters
JobFilter · object[]