How the Job Pool Works
Four sources feed this page: JSearch (which wraps Google Jobs), Remotive, Adzuna, and Wellfound. Ingestion cadence is tiered by how actively a search is being used, from twice a day down to every few days for dormant criteria. Each incoming listing gets deduplicated by title + company + location hash, then queued for scoring. The search bar queries the full pool. Results update on keystroke. The 20 listings you see on first load are the most recently posted active jobs.
What Happens Before a Job Reaches This Page
Every listing passes through a processing pipeline. First, the apply URL is resolved: the engine follows redirect chains up to 8 hops, strips UTM and tracking parameters, and checks that the destination returns a 200. If it hits a 404 or a "this position has been filled" page, the job gets flagged as closed. Salary data is extracted when the posting includes it, estimated from comparable postings when it does not (those show a "~" prefix). Ghost probability is calculated from posting age, repost frequency, and description specificity.
Filters and Search
The remote toggle restricts results to postings explicitly marked as fully remote across all 5 sources. Remotive is the densest source for distributed roles, but Adzuna and JSearch carry remote listings from larger employers. Search matches against title, company name, and description keywords. Without an account, you get the first 500 characters of each description and 20 results per query. Logged-in users see full descriptions and up to 100 results.
Frequently Asked Questions
Which job boards feed this page?
JSearch (wraps Google Jobs), Remotive, Adzuna, and Wellfound. Refresh cadence is tiered by how actively a search is being used, from twice a day down to every few days for dormant criteria. Every incoming URL is resolved, cleaned of tracking parameters, and verified to return a live page before appearing here.
How does the scoring engine generate a fit score?
The engine parses the job description, extracts every requirement (skills at specific depth levels, years of experience, location constraints, degree preferences), and compares each one against your profile. The result is a 0 to 100 score with a per-requirement breakdown showing matches and gaps. You need a profile with a resume uploaded for this to run.
How are ghost jobs and dead listings caught?
Three layers. First, HTTP status: a 404 or 410 means the page is gone. Second, redirect analysis: if the URL resolves to a generic "position not found" landing page, it is flagged. Third, page body scan for expired-listing language ("this role has been filled," "no longer accepting applications"). On top of that, a ghost probability score factors in posting age, repost count, and description specificity. Flagged listings are marked closed and removed from active results. The detection methodology page covers each layer in detail.
Is there a remote filter?
Do I need an account?
No. Search and browse work without one. Anonymous access is capped at 20 results per query and 500-character description previews. An account raises the limit to 100 results, shows full descriptions, and enables scoring.
How do salary estimates work when the posting omits pay?
The engine matches by title, location, remote status, and seniority against postings that do disclose salary, then produces an estimate. Estimates are prefixed with "~" so they are visually distinct from disclosed ranges. Accuracy improves as the salary corpus grows. The Role Intelligence pages show aggregate salary data (median and average) across all disclosing postings for a given title.