Job Radar — Deduped, Ghost-Filtered Job Alerts

Fresh, real job openings — duplicates merged, stale and ghost-flagged postings pushed down or hidden, every job ranked for you.

Job boards have two dirty secrets. First, the same opening is syndicated everywhere — one job has been found copied across 200+ boards and feeds, so half of what a multi-board search returns is the same role wearing different URLs. Second, a lot of what's posted is never going to hire anyone: 18–22% of postings are "ghost jobs" (Greenhouse, 2025), 4 in 10 hiring managers admit posting a fake one in the past year (ResumeBuilder), and hires-per-posting have halved since 2019 (Revelio Labs). Scrapers happily hand you all of it, duplicated and stale.

Job Radar is the finished product instead of the raw feed. Type what you're looking for and where, add your email if you want the digest — and every run brings back one clean, ranked list: duplicates merged across boards, every job scored 0–100 for how well it matches you and checked against a ghost-job checklist, with the apply link that goes to the employer's own site whenever one exists. Run it on a Schedule and it becomes a radar: after the first run you only hear about what's new.

A note on the ghost filter, since the number above is about the market and not about your results: the score is deliberately conservative. Only the clearest cases are hidden outright, and on a fresh daily digest that is often none — a job posted today isn't a ghost yet. What the score does on every run is rank suspicious postings down and show you the reasons on the card ("no salary listed", "not found on the employer's careers site", "only reachable through a job-aggregator repost"). It earns its keep on wider searches, where older and recycled listings actually turn up. We'd rather show you a stale posting with its reasons attached than quietly bury a real job to make a statistic look better.

No auto-apply. Job Radar finds and ranks; you read and click. Your applications stay yours.

Priced lean — pay only when you run, no subscription. Missing a board, a country, a filter? Open an issue on the Issues tab — we read every one. 🙏

Where it searches

Source	What it covers	Powered by	Default
Companies you follow	Greenhouse / Lever / Ashby career pages — direct from the employer, free	built in	✅ prefilled
Google for Jobs	20+ boards in one sweep (the spine), each with a real apply link	johnvc/Google-Jobs-Scraper	off — opt in
Indeed	the biggest single board, very cheap	valig/indeed-jobs-scraper	off — opt in
Glassdoor	salary-rich listings	valig/glassdoor-jobs-scraper	off
LinkedIn	public listings only — no login, no cookies	cheap_scraper/linkedin-job-scraper	off

By default Job Radar runs the free career-page fetcher only — zero setup, no scraper credit, results in seconds. The board scrapers (Google for Jobs, Indeed, Glassdoor, LinkedIn) run as pay-as-you-go Store Actors on your own Apify account (cents per run — see Costs); tick the ones you want. The career-page fetcher is ours and costs nothing: add greenhouse:stripe, lever:palantir, ashby:ramp — or just the company slug and we probe all three.

The quality layer (what you're actually paying for)

• One row per real job. We normalize company, title, seniority and location, then merge duplicates across boards — the canonical record keeps the best apply link (employer ATS first), the earliest true posting date, and lists every other board it appeared on. Being on 3+ boards is itself a legitimacy signal, and we show it.
• Ghost & stale postings are scored, shown, and only then hidden. Every job gets a 0–100 ghost score from signals like true age (median time-to-fill is ~44 days — a posting alive past 90 is a red flag), repost churn, "always hiring" evergreen language, missing salary where the law requires one, one company blanketing 10+ cities with the same title, and aggregator-only listings that don't exist on the employer's site. The score always pushes a suspicious posting down the ranking and prints its reasons on the card. Only a job scoring 70+ is hidden outright — a high bar, so on a fresh daily digest the "we hid N" counter is often 0 ghosts and the hidden count is mostly off-target titles. That's deliberate: a false "ghost" verdict costs you a real job.
• Repost-proof age. Boards let companies repost to reset the "posted 2 days ago" clock. We keep our own first seen date across runs — their clock resets, ours doesn't.
• Exact matches, not keyword soup. A search is a phrase, not a bag of words, and it is matched against the title only. A one- or two-word search matches whole titles only: react developer returns jobs titled React Developer, and deliberately not Sr React Developer, Front-End React Developer or React Developer Intern — at that length a substring is a fragment of half the market. From three words up, a title that contains your phrase counts too, because by then the phrase is the role: senior react developer also catches Senior React Developer, Payments. Want the variants? Give each one its own line — or, in a thin market where only one board answers, switch Title match strictness to Broad to also keep titles that carry all your words in any order (so frontend developer catches Medior Frontend Developer). Everything off-target is counted in the "we hid N" panel, never silently dropped.
• An explainable match score. Each job is scored 0–100 against your searches, must-haves, salary floor and seniority — and shows which of your searches it matched, not a black-box number.
• Only what's new. The first run is your baseline; after that, each run flags what's new, what changed (salary added, reposted), and what disappeared — perfect on a daily or weekly Schedule.

What you get

• 🔗 A shareable dashboard page — every run publishes a clean, branded page to your own Apify storage: every search it ran with its match count, the ranked list with apply buttons, filter chips (new · remote · salary listed · direct employer) plus a filter per search, sorting, and the hidden-postings counter with one-click reveal. One link, no login, dark & light mode, works on your phone.
• 📧 An email digest — type an address and each run mails you the new matches first (subject like "Job Radar: 12 new matches for react developer — 3 hot, 14 hidden as stale"), plus what changed and what closed. Remove the address to stop.
• A clean dataset — one row per unique job (title, company, location, salary, apply link + type, posted date, match score, which of your searches it matched, ghost score with reasons, which boards carried it) for your own spreadsheets and automations.

Set it up in 2 minutes

1. Enter the job titles you want — one per line, matched against the title (Senior React Developer, then React Engineer if you want both) — and a location (or flip Remote only). Matching is Strict by default; switch Title match strictness to Broad for thin markets.
2. (Optional) Add must-have / exclude keywords, a salary floor, seniority — and your email.
3. Companies you follow comes prefilled with a few examples for free direct career-page coverage — edit the list to track your own.
4. Hit Start. Then use the Schedule tab to make it daily — that's when "new since last run" starts working for you.

Costs

One flat fee, only when you run — no subscription, never per job.

• $0.10 per GB of run memory at start — at the default 1 GB that's $0.10 per run, roughly $3/month as a daily digest. Dedup, ghost filtering, exact-match ranking, the dashboard and the email are all included.
• The optional board scrapers bill their own pay-as-you-go fees directly to your own Apify account, and we never mark them up. The default run uses the free career pages only, so it adds no scraper credit — just the flat fee. Google for Jobs and Indeed are both very cheap (fractions of a cent per listing), so enabling them adds only a few cents of scraper credit per single-search run. Max matching jobs per source is a post-filter target — the number of matching jobs kept per source; under the hood we fetch up to 3× that many raw rows (capped at 300) and whittle them down, so lowering it still caps what the scrapers bill you. Each extra search line and each optional source (Glassdoor, LinkedIn — LinkedIn bills a 150-result minimum when on) adds proportionally.

Ask your AI assistant (MCP)

Works as an agent tool out of the box: connect mcp.apify.com to Claude, Cursor, or any MCP client and just ask — "find me senior React jobs in Berlin" or "anything new for my job search today?" — and it runs this Actor with the right inputs.

Notes

Reads publicly visible listings only and logs into nothing — LinkedIn coverage is public-listings-only, off by default, and skipped gracefully if unavailable. The board scrapers run on your own Apify account (your token — see the input's advanced field); your job list, history and dashboard live in your own Apify storage. We host nothing, and we never apply on your behalf.