ZipRecruiter Job Search

by agents.openwebninja.com·Website ↗API ↗

ZipRecruiter Job Search is a paid API for AI agents from agents.openwebninja.com, paid per call via MPP, price unknown, status unknown (last checked 2026-10-14).

Search ZipRecruiter job postings by keyword with server-side filters for location, salary, employment type, experience level, and more, with optional full job detail enrichment.

How this service works

Search ZipRecruiter job postings by keyword, with the filters ZipRecruiter honors server-side: `location` with a kilometre `radius`, `date_posted`, `employment_types`, `experience_level`, `salary_min` / `salary_max`, `remote_jobs_only`, `zipapply_only` and `company_id`. **`include_details` defaults to `false`, and this is the single most important thing to understand about this endpoint.** A ZipRecruiter search card carries a roughly 200-character teaser and nothing more. `job_description`, `job_highlights`, `employer_website`, `employer_reviews`, `job_onet_soc`, `ziprecruiter_description_html`, `ziprecruiter_job_industry`, `ziprecruiter_valid_through`, `ziprecruiter_ats_identifier`, `ziprecruiter_company_size`, `ziprecruiter_company_hq`, `ziprecruiter_company_industries`, `ziprecruiter_key_responsibilities` and `ziprecruiter_employer_faq` are all detail-only and come back `null` on a plain search. Set `include_details=true` to fill them. The teaser is always published separately as `ziprecruiter_short_description`, so a card-only response still carries usable copy. Filling the details costs one extra upstream request per job, which is why it is opt-in. With `include_details=true`, `num_pages` may be at most 5, which is 100 jobs; a larger value returns a 400 error. **Pagination** is 20 jobs per page. The ceiling is `ceil(total_results / 20)`, where `total_results` is ZipRecruiter's own count of matches, and the largest result set observed was 946 jobs. Duplicates ZipRecruiter repeats across adjacent pages are removed, so a page may yield fewer than 20 new jobs. The response is a flat array, so a call returning fewer than `num_pages` x 20 jobs is how you know you have reached the end of the result set. **Scoping to one company** is what `company_id` is for, and the value is the `ziprecruiter_company_id` carried on every job. ZipRecruiter fails this filter OPEN in two ways, both silent: a bad id returns the full unfiltered search, and a valid one intermittently drops the filter. Send `location` alongside `company_id`: that is what keeps the drop rare. Measured on one id with five interleaved repeats per arm, the filter held 5/5 with a location and 1/5 without, and the misses returned 6, 4 and 20 distinct employers. So every returned card is checked against the requested id, a page that fails the check is refetched, and an id that stays unscoped returns a 404 error. `total_results` is omitted for a company-scoped search: ZipRecruiter's count there swings from 1 to 540 across 20 identical calls. Note that roughly a third of postings carry the placeholder id `zr_missing_company_default`, ZipRecruiter's marker for a posting whose company record is missing; it returns a 404 error here, so pick an id off a job whose `employer_name` is the company you want. **Sorting.** Results come back in ZipRecruiter's own relevance order. `sort_by`, `sort`, `order` and a `company` NAME each return a 400 error, because ZipRecruiter accepts and then ignores them. **Billing:** `Requests` counts one per requested SERP page, plus one per detail page when `include_details=true`. Extra pages walked to refill a page after deduplication, and retries of a blocked page, are absorbed.

Output

Returns a paginated array of job postings from ZipRecruiter. Each result includes job ID, title, employer name, location, employment type, salary info, apply link, and a short teaser description. When include_details=true, also returns full job description, highlights, employer website, employer reviews, company size, company HQ, key responsibilities, valid through date, ATS identifier, and employer FAQ. Also includes total_results count and request metadata.