Page Size now goes up to 10,000, and Max Results leaves the input form. You pick the page and how many jobs per page, as before: page 1 with Page Size 1,000 returns the best 1,000 matches in one run, and page 2 returns the next 1,000. Pages above 100 jobs are fetched 100 at a time, so one run returns up to 10,000 jobs; it stops early when fewer jobs match or when it reaches the run's maximum cost. Inputs that set Max Results in 2.12 keep working. Pricing is unchanged.
New Max Results input: set how many jobs you want, up to 10,000, and one run returns them, best matches first. Before, a run returned at most one page of 100. The run stops early when fewer jobs match or when it reaches the run's maximum cost. Page Number and Page Size leave the input form but still work in JSON input and API calls, and a run without Max Results returns the same single page as before, so existing runs and schedules are unchanged.
Company filters now tell you when an entry did nothing. Before, a Companies (Exclude) entry that matched no company, such as "Lockheed" for Lockheed Martin, excluded nothing and the run gave no sign of it. Now the run log and the run's status message name every unmatched entry in either company filter and suggest the company it probably meant, e.g. Did you mean "Lockheed Martin" (lockheedmartin.com)?, and a company that hires on a different ATS is named along with the actor that covers it. Several names pasted into one entry ("Acme, Globex") are now split into separate companies, so an entry like that starts working; before, it matched nothing. The Company match report in the Output tab, which earlier versions listed but never wrote, now records the result for every run that uses a company filter. Pricing is unchanged.
New Description Format input. HTML returns each job's description as HTML instead of Markdown. Enhanced Markdown (Beta) and Enhanced HTML (Beta) return the same words with a consistent layout: real headings, lists, key facts as **Label:** value, and the repeated title, buttons and tracking tags removed. The layout is produced automatically and is not guaranteed to be perfect; a job without an enhanced description returns its standard one. The default stays Markdown, so existing runs and schedules are unchanged.
Jobs can now carry a brand object: the brand the hiring company trades under when it is part of one, such as a franchisee, an independent agent or agency, or a hotel run under a hotel group's flag. company stays the employer that posted the job; brand is the brand customers know it by (e.g. Culver's), with its id, name, domain and logo_url. For hotels it is the hotel group, not the individual flag. The key is omitted on jobs whose company is not part of a brand. The output table and the actor's input are unchanged.
When a run finishes, the actor now tells Jobo its run id, your Apify user id, whether you are on a paid Apify plan, how many results it was charged for and the total charged, so we can track our Store revenue. Your input and the jobs it returned are not sent. Results and pricing are unchanged.
Jobs now carry an occupation object: what kind of job it is, as a category › subcategory › occupation (e.g. Healthcare › Nursing › Registered Nurse) with its id, ISCO-08 code and O*NET-SOC codes. It is omitted when a job is not classified. The output table gains Occupation and Job category columns. New inputs filter on it: Job Category and Job Subcategory (pick from the list) and Occupations by id (copy occupation.id from any result).
The full company profile (include_company_details) no longer carries personal data or news: founders, leadership, phone_number, email_address and press_references are removed, along with the Founders and Leaders columns in the output table.
Company filters now explain themselves. The run log, the run status message, and a
new COMPANY_MATCH_REPORT output record show which of your companies_include
names matched, which did not, and which are indexed on a different ATS. Zero-result
runs name the cause instead of "No jobs found matching criteria".
The locations field is no longer pre-filled — it silently restricted every run to
a few example cities, which was the most common reason a correct company name
returned nothing.
- Maintenance: input-schema cleanup.
- Added
posted_before for bounded posting-date windows.
- Enforced a minimum
page_size of 1.
- The full
description field now appears directly in the job-listings view.
- Company enrichment is opt-in:
include_company_details defaults to off; jobs carry a lightweight company preview.
- Opt-in company enrichment:
include_company_details replaces the company preview with the full profile (funding, leadership, ratings, tech stack, H1B history), fetched once per unique company.
- New "Company details" output view.
- Richer job shape: AI-extracted
qualifications (typed hard/soft skills, education, certifications), responsibilities, benefits, summary, and normalized_title.
- Geocoded
locations (city / region / country / latitude / longitude).
- New filters: salary range, skills include/exclude, companies include/exclude, work model, experience level, employment type, and negative query syntax.
- Work-authorization signals:
is_work_auth_required, is_h1b_sponsor, is_clearance_required.