Oortcloud
Email us
Oortcloud Job HunterJob-market intelligence

Every role on a thousand career pages, read for you every night.

Job Hunter visits the live career page of every company you track — hundreds or thousands of them — deduplicates what it finds, and emails you only the roles that are genuinely new since the last run.

> 70+ ATS scrapers live · no cached layer · delta-only alerts

Email us about accessHow it works

Used today by financial services and travel recruitment teams, executive search consultancies, and in-house talent intelligence functions.

Extract · 20 Oct 06:00 UTClive scrape
512
Companies tracked
12
New since last run
1,438
Still open
9
Closed this run
Head of Risk Analytics4 days open
Meridian Bankats: Workdayown site
Senior Pricing Analyst2 days open
Northline Airats: Greenhouseown site
Data Platform Lead9 days open
Kestrel Asset Mgmtats: Leverown site
Director, Operational Resilience31 days open
Calder Insuranceats: SuccessFactorsown site
3career_blocked
14no open roles
0page unreachable
1filter emptied set

Illustrative extract — sample data

01The problem

A 500-company patch is normal. Checking it by hand is not.

Doing this properly by hand is not a resourcing problem you can hire your way out of. It is a throughput problem.

01

Hundreds of career pages, every morning

A patch of 200, 500 or 1,000 target companies is normal at mid-market and executive-search level. Opening those pages by hand is a non-starter, and most teams cannot spare a researcher to do it.

02

Aggregators are stale, incomplete and noisy

Roles routinely appear on a company's own career page days before they surface on a job board — and many never appear on a board at all. LinkedIn, Indeed and the sector boards are a lagging sample of the market, not a record of it.

03

No way to see what is new today

Without a stored previous snapshot there is no way to tell what appeared overnight from what was already there. The value is in the delta, and a list of open roles does not have one.

04

Hiring signals stay invisible

A new team being built, an unusual spike in mid-level roles, a sudden cluster of restructuring hires — none of it is visible unless you can see the whole picture across dozens of companies at once.

05

The same report, rebuilt by hand

Sharing findings with a colleague or a client means rebuilding the same report manually, repeatedly, with no audit trail and no way back to the run it came from.

Job Hunter is built on one principle: the company's own career page is the source of truth — and the recruiter should not be the one sitting there reading them every morning.

02Core capabilities

What the platform actually does, end to end.

Eight moving parts that turn hundreds of career pages into a single, trustworthy picture of the market.

01

Live career-page scraping

Every search visits the live career page of each selected company at the moment the search runs. There is no cached layer serving old data, so the result is a faithful snapshot of what the company is posting right now.

  • Static HTML parsed with Cheerio — fast and low-overhead
  • JavaScript-rendered pages driven by headless Chromium through a shared browser pool, so React, Angular and Vue career sites load fully before extraction
  • ATS-embedded feeds handled by dedicated parsers rather than guessed at
  • Anti-bot detection and bypass for Cloudflare managed challenges, AWS WAF, Incapsula, Akamai and PerimeterX, falling back to stealth-browser mode
  • Sites that stay unresolvable — interactive captchas, Turnstile — are flagged career_blocked rather than silently returning zero
02

70+ dedicated ATS scrapers

Purpose-built parsers for the world's major recruiting platforms, each handling that vendor's URL conventions, JSON endpoints, pagination quirks and session requirements.

  • Enterprise: Workday, Oracle Cloud HCM, SuccessFactors, Dayforce, iCIMS, BrassRing/Kenexa, Avature, PageUp, Phenom, Eightfold
  • Mid-market and modern SaaS: Greenhouse, Lever, Ashby, Workable, SmartRecruiters, Personio, Teamtailor, Recruitee
  • UK and Europe specialists: Eploy, Oleeo, iTrent, Hireserve, Networx, CivilServiceJobsUK, eArcu, AllHires
  • Agency and SMB: Bullhorn, Vincere, Loxo, JobAdder, ZohoRecruit, Radancy/TalentBrew
  • Unrecognised or bespoke platforms fall through to a generic multi-tier scraper, so coverage degrades gracefully instead of failing silently
03

Intelligent deduplication

The same role on the company's own site, on LinkedIn and on a sector board surfaces once — not three times — with every source listed against it.

  • Exact URL matching, catching identical links across sources
  • Fuzzy title and company matching, catching minor wording variants of one role
  • Semantic matching over OpenAI embeddings in pgvector, so a Head of Risk and a Risk Management Lead at the same company resolve to one brief
  • Source precedence: the company's own career page or ATS is always the master — board data supplements it but never overrides it
04

Role lifecycle tracking

Every role is tracked from first appearance to removal, so you can answer how long a brief has been open, not just whether it is open.

  • First seen date — when the system first discovered the role
  • Days open, continuously updated
  • Removal detected only after two consecutive missed scrapes, with cliff-guard protection so one bad scrape cannot wipe a company's role set
  • Fast close when a scrape is clearly complete and confident, cutting latency on genuine closures
  • Repost detection: a role removed and later relisted surfaces its repost history in analytics
05

Four ways to search

Company Search, Title Search, AI Search and Match Roles — because different questions deserve different tools.

  • Company Search: pull every open role across a chosen subset of tracked companies, with title filters, in live-scrape or saved-data mode
  • Title Search: one job title across every tracked company and configured board at once. Saved Data Only returns instantly at no scraping cost
  • AI Search: describe the brief in plain English — “Chief Risk Officer roles at UK-listed banks” — against your own OpenAI key
  • Match Roles: paste a role description or upload a CV and rank every open role in your universe by calibrated match percentage, with a title-only fallback so nothing is invisible
06

Saved searches and batch scheduling

A saved search is a named configuration of companies, filters and recipients — run it when you want, or let it run overnight.

  • Nightly automated batches that compare results against the previous run
  • Per-search recipient lists, so different searches notify different people
  • Run Batch Now for an off-schedule refresh
  • Queue-based scheduler with concurrency limiting, so scheduled searches do not collide
  • Background auto-refresh re-scrapes tracked companies that have not been visited recently — nights and weekends included
07

Delta-only email alerts

Recipients hear about the twelve roles that appeared since yesterday, not the same four hundred open roles every morning.

  • Role title, company, location and a direct link to the listing
  • First seen date and the source it came from — company site, LinkedIn or a sector board
  • Filter context, so a recipient knows why they are being told
  • Delivered through MailerSend with per-organisation sender configuration
08

Immutable extract archive

Every search — interactive, batch or AI — produces a timestamped extract stored in the database. What was open on a given date stays answerable.

  • A snapshot of every role exactly as at the moment of the search
  • Breakdown by company: how many roles, which scrapers found them, what went wrong
  • Searchable, filterable and shareable after the fact
  • Title filters recorded alongside the results they produced
  • Downloadable as CSV for a client, a business partner or your own records
03ATS coverage

70+ dedicated scrapers — maintained, not guessed at.

Generic scrapers fail on ATS-hosted career pages because every vendor has its own URL conventions, JSON endpoints and pagination. Each platform below has a purpose-built parser.

Enterprise

WorkdayOracle Cloud HCMTaleo ClassicOracle RecruitingSuccessFactors RMKSuccessFactors UnifySuccessFactors j2wDayforceiCIMSBrassRing / KenexaAvaturePageUpPhenomEightfoldFindly CTS SymphonyTalent-CommunityTalentSoftSuccessFactors Marketing-Front

Mid-market and modern SaaS

GreenhouseLeverAshbyWorkableSmartRecruitersBambooHRPersonioHarriParadox / OliviaOccupopBizneo HRCareer Site CloudTeamtailorBreezy HRRecruiteeJazzHRComeetJobviteTalemetryRipplingClearCompanyADP Workforce NowPaycomPaylocity

UK and Europe specialists

EployOleeo (tal.net)iTrentHireserveNetworxPeopleHRTalos360BeeSiteCivilServiceJobsUKPageUp cw/ci skinseArcuNLX / DirectEmployersAllHiresiApplyApplaud

Agency and SMB

BullhornVincereLoxoJobAdderZohoRecruitATS OnDemandChetuRadancy / TalentBrewJobTrain

Job boards and networks

LinkedIneFinancial CareersSmartMoneyMatchStudySmarterCustom configured boards

Bespoke platforms and unrecognised ATSs

Anything running an unrecognised or bespoke platform falls through to a generic multi-tier scraper — HTTP, then stealth browser, then structured extraction — so coverage degrades gracefully rather than failing silently. Sites that stay behind an interactive challenge are reported as career_blocked instead of being quietly counted as empty.

04How it works

Define what to watch and when. Job Hunter does the looking, the comparison and the analysis.

Results are deduplicated across sources, classified and stored as a snapshot per company. The notification layer compares the latest snapshot with the previous one to identify what is genuinely new. Every search produces an immutable extract that can be revisited, shared or exported — while background workers precompute the analytics so dashboards load instantly.

01

Track a universe

Give each company one or more career-page URLs. Organisations typically run from hundreds to low thousands of tracked companies.

02

Define a saved search

A named group of companies, optional title filters, and the colleagues or clients who should be notified.

03

Run it, or schedule it

Interactive on demand, nightly as a batch, or immediately with Run Batch Now. Background refresh keeps the rest of the universe current.

04

Fan out across scrapers

One engine routes each URL to its ATS-specific parser, falling back to structured HTML extraction or JavaScript rendering for unrecognised platforms.

05

Deduplicate and store

Results are deduplicated across sources, classified as own-site, job-board or AI-sourced, and stored as a per-company snapshot.

06

Diff and notify

The latest snapshot is compared with the previous one and only genuinely new roles are emailed. Analytics workers keep trends and BD signals precomputed.

05Who it's for

Built for the people who carry the target list.

Executive search consultants
Map a target universe of 50–500 companies and get alerted the moment a relevant brief opens — before the company has briefed any agency.
Contingency recruiters
Track a defined sector patch continuously, and reach candidates with live roles rather than stale intelligence.
In-house TA strategy and talent intelligence
Monitor competitors' hiring patterns, benchmark headcount movements and build internal market maps for business partners.
PE and VC operating partners
Use hiring velocity as a leading indicator of operational health, expansion or stress across portfolio and target companies.
Business development teams
Identify companies experiencing hiring pain — overdue roles, reposting patterns, unusual volume spikes — as warm BD triggers.
Research and competitive-intelligence analysts
Turn hiring signals into leading indicators of product direction, geographic expansion, leadership transitions and org restructures.
Sector-focused agencies
Financial services and travel are the primary specialisms today; the engine is industry-agnostic and can be pointed at any defined universe.
06Analytics and intelligence

Raw listings become a picture of the market.

Analytics are computed from the system's own snapshot history — no third-party data dependency, and nothing that depends on a board deciding to tell you what changed.

01

Role Analytics

An organisation-scoped dashboard computed from your own snapshot history — no third-party data dependency.

  • Aged roles — which companies have roles open 30, 60 and 90+ days, sortable by company or role
  • Company hiring trends — opens versus closes, net hiring, median time-to-close and stuck roles
  • Peer ranking by current open-role count, growth rate and sector
  • Geography — where companies are hiring, by country and city
  • Source mix — the proportion of roles found direct from career pages versus job boards
  • Scrape momentum and coverage — what is current, and what is overdue
02

Market Intelligence

Sector-level trend analysis across the full tracked universe.

  • Emerging titles gaining share over the trailing period
  • Hiring velocity — which companies are accelerating and which are stalling
  • Competitor demand mapping across a defined peer group
  • AI-generated company profiles describing current hiring posture
03

BD Call Sheet

The business development view: every tracked company scored on a 0–100 opportunity scale and tiered, then downloadable as a dated PDF.

  • Stuck-book share — open roles overdue beyond 45 days, implying briefs a recruiter could take on
  • Repost pain — roles relisted after closing, implying placement failure or a re-opened brief
  • Competitor pressure — sector peers increasing volume while this company is flat or shrinking
  • Expansion signals — new geography, new team, net-new titles appearing
  • Footprint breadth — hiring across multiple functions at once
  • Slowdown signals — recent deceleration suggesting a company that may need to rethink
04

Intelligence Reports

AI-generated company reports, produced on your own OpenAI key, stored and shared across the organisation, regenerable on demand.

  • Organisational structure and key hires
  • Current hiring posture and open-role mix
  • Headcount trajectory and hiring forecasts
  • Suggested talking points for a BD or candidate call
05

Executive Search

An asynchronous scan for senior leadership and board-level opportunity, delivered as a printable PDF report.

  • Searches for real-world trigger events: restructurings, M&A activity, PE investment, leadership transitions and geographic expansion
  • Inputs: industry or sector, optional sub-sector, optional locations
  • Output: executive summary, ranked opportunity table, company profiles with trigger events and signal strength, recommended BD actions and a sources index
  • Completed reports are shared across the organisation

BD Call Sheet — signals and tiers

Each company is tiered from its score, and freshness-gated: a company that has not been scraped recently is never surfaced as Hot — it is marked Refresh First so you know to re-run before you call. Filter cards slice the sheet instantly by signal type.

Tiers

HotWarmWatchRefresh FirstQuiet

Filter cards

Call TodayRefresh FirstRepost PainAged BacklogExpansion AlertsEnterprise MappingNiche Specialist
07What makes it different

Direct from the source, and honest about what it cannot reach.

Source of truth, not a feed

Most products in this space repackage LinkedIn, Indeed or another aggregator. Job Hunter scrapes the company site directly, so you see roles that never reach a job board — and you see them earlier.

70+ ATS parsers, maintained

Generic scrapers fail on ATS-hosted career pages because every vendor has its own URL structure, JSON endpoint and pagination convention. A Workday company and a Greenhouse company both return complete, structured results — not the fraction visible in the initial HTML.

Anti-bot resilience, not workarounds

Many career pages sit behind Cloudflare, AWS WAF or another challenge. A tiered strategy — stealth browser, cookie harvesting, challenge solving including Altcha proof-of-work — instead of returning zero when a challenge fires.

Role lifecycle, not just a list

Knowing a role is open today is useful. Knowing it has been open 73 days, was reposted twice, and that three similar roles opened at competitor companies is intelligence.

Honest failure modes

A zero result can mean no open roles, a blocked page, or a filter that removed everything. Job Hunter tells you which one it is. There is no ambiguous silence.

Team workflow built in

Organisation-scoped data means shared company universes, shared saved searches, shared analytics and shared notifications from day one — no personal-list-versus-team-list problem to solve later.

A cost model you control

AI features run against the organisation's own OpenAI and Exa keys. Nothing is put on a shared meter and marked up.

08Why not just use ChatGPT or Claude?

A language engine is not a data system.

It is the first question anyone asks. The honest answer is that a general-purpose model cannot reliably do this work, and a thin wrapper around one is much further from it than it looks. Here is everything that would have to be bolted on — and what breaks if any one of it is missing.

A real fetching layer

A proper HTTP client plus a JavaScript-rendering headless browser, because most modern career pages do not ship server-rendered HTML. A browse tool doing a single GET per query does not render JavaScript at scale.

A library of ATS-specific parsers

Greenhouse, Lever, Workday, Ashby, SmartRecruiters, iCIMS, Taleo, SuccessFactors, Phenom, PageUp and sixty more each have their own URL conventions, JSON endpoints and pagination rules. Asked to find the roles on a company's site, an LLM guesses at the structure and silently mis-extracts.

State between runs

An LLM has no memory of last Tuesday. The entire notification value depends on a per-company, day-by-day snapshot history in a database. Without it, there is no such thing as new.

Real deduplication

One role appearing on the company site, LinkedIn and a sector board gets listed three times. Doing it properly needs normalised titles, fuzzy matching, semantic comparison and a stable per-role identity across runs.

Anti-bot survival

Cloudflare, AWS WAF, Incapsula, Datadome, IP rate limits, JavaScript challenges and cookie walls. An LLM hitting these returns a generic failure — or hallucinates results from the block page.

Reliability and reproducibility

Every Job Hunter run produces an immutable, timestamped extract you can revisit, share, audit and export as CSV. LLM output is non-deterministic: ask the same question twice and you get two different answers, neither timestamped nor stored.

Honest failure modes

Job Hunter separates no open roles from a blocked career page, from no career URL on file, from a filter that wiped everything out. An LLM tends to invent a plausible answer rather than report that it could not reach a page.

Scale economics

Tracking 500 companies means 500+ page fetches per run, often with several URLs each, often needing JavaScript rendering and challenge solving. That is impractical and expensive through a chat interface.

Team workflow

Multi-user organisations, shared tracked-company lists, saved searches, scheduled batches, alerts to specific recipients, role-based access and audit logging. None of that lives in an LLM.

Domain-specific filtering

Career pages are full of text that looks like a job title to a language model: “Why work here?”, “Vice President” with no further context, “Do you have a passion for travel?”, navigation links labelled Careers. Without rules, these are listed as roles and trust in the data collapses immediately.

Where AI genuinely helps, we use it

Company intelligence reportsExecutive search scansOrganisational summariesHiring forecastsNatural-language and semantic search

All of it sits on top of a system that knows the difference between a real role and a navigation link, that gives you the same answer twice when nothing has changed, and that remembers what was open last Tuesday.

09Data integrity and reliability

Authentic data, always — and no ambiguous silence.

Every role in the system traces back to a live scrape. Use Saved Data Only queries the stored snapshot from the last scrape, which is timestamped: there is no mechanism for serving stale data as fresh. When something cannot be reached, the system says so.

No open roles

The career page was reached and is genuinely empty.

Career page blocked

An anti-bot wall was detected and not cleared, so the company is flagged career_blocked rather than reported as hiring nothing.

Page unreachable

A network or DNS failure stopped the request before any content was returned.

Filter removed everything

The search had title filters applied that matched nothing in the live set.

No silent role deletion

A role is marked closed only after two consecutive missed scrapes — never after a single bad scrape or a temporary block. Cliff-guard logic detects suspicious mass drops and preserves prior data pending re-confirmation, while fast-close logic accelerates genuine closures when a scrape is clearly healthy.

Source-guard

If a scraper that previously found hundreds of roles suddenly returns zero, that result is treated as a failure and the prior snapshot is preserved — rather than mass-deleting the company's entire role set.

Immutable history

Snapshot history is never overwritten. You can always see what was open on any date for which a scrape was run.

10Team and organisation

One shared picture of the market, from day one.

There is no personal list versus team list problem to solve later. Tracked companies, snapshots, searches and analytics belong to the organisation.

Organisation-scoped data

Tracked companies, snapshots, saved searches, notifications, documents and analytics are shared automatically inside an organisation. Everyone sees the same picture, and nobody re-scrapes for their own view.

Role-based access

Admin, standard user and viewer roles. Admins manage tracked companies, API keys and organisation settings; viewers have read-only access.

Per-organisation API keys

OpenAI and Exa keys are stored per organisation, so every AI feature — intelligence reports, executive search, AI search, semantic matching — runs against your own key. No shared deployment-wide spend, no cost surprises.

Company document store

Attach PDFs, Word documents, spreadsheets or notes to any tracked company: briefs, org charts, candidate longlists, previous proposals.

Audit trail

System events — scrapes, notifications, report generation, user logins — are logged with timestamps for accountability.

11Specialisms

Deep today in financial services and travel; pointed anywhere tomorrow.

The system is deployed with a strong specialisation in two markets, but the scraping engine, analytics and intelligence features are fully sector-agnostic.

Financial services

Banking, asset management, wealth management, insurance, FinTech and payment infrastructure.

Travel

Airlines, tour operators, travel technology and hospitality.

What the specialism gives you

  • Pre-built tracked-company lists covering hundreds of organisations in each specialism
  • Sector-aware analytics presets: peer-group comparisons, typical time-to-fill benchmarks and sector-specific title families
  • BD Call Sheet signals tuned to recruiting-firm patterns in those markets

Point it at anything

Professional servicesTechnologyHealthcareConsumerAny defined universe of companies
12At a glance

The specification, in one place.

Used by
Financial services and travel recruitment teams, executive search, in-house TA intelligence
Primary sector specialisms
Financial services and travel — the engine is industry-agnostic
Tracked companies
Hundreds to low thousands per organisation
ATS coverage
70+ dedicated scrapers, with a generic fallback for unrecognised platforms
Search modes
Company Search · Title Search · AI Search · Semantic Match Roles
Scheduling
On-demand interactive, nightly batch and background auto-refresh
Analytics
Role lifecycle · Hiring trends · Peer ranking · Geography · Source mix · BD scoring
Intelligence
AI company reports · Executive search · BD Call Sheet PDF
Hosting
Replit Cloud — managed PostgreSQL via Neon, object storage, email delivery
Data freshness
Live at scrape time; no cached layer
Privacy model
Per-organisation data isolation; per-organisation API keys
AI integrations
OpenAI for intelligence reports, executive search, embeddings and AI search; Exa for AI Search web lookup
13Get started

Map your universe.

Tell us the sector you cover and the companies you want watched, and we will help you set up your first tracked universe. Everything is configured around your patch — no generic demo data.

Prefer your own mail client?

support@oortcloud.ai

Send an enquiry

  • The sector and locations you cover
  • Roughly how many companies you want watched
  • Who should receive the new-role alerts