Multi-ATS Tech Hiring Signals

Detect when companies start or stop investing in SRE, DevOps, platform engineering, cloud infrastructure, security, AI infrastructure, data engineering, observability, FinOps, and quantitative engineering.

This Actor reads the public, unauthenticated job-board APIs published by Greenhouse, Lever, and Ashby. It stores a bounded snapshot for each watchlist, compares scheduled runs, and emits explainable company-level changes with direct source links. It does not log in, scrape profiles, collect applicant data, bypass access controls, or invent contact information.

Why buyers use it

• Cloud, observability, security, and developer-tool vendors can prioritize accounts that are actively building relevant teams.
• Recruiters can monitor target employers without repeatedly downloading the same job list.
• Analysts and AI agents receive compact, evidence-linked signals instead of thousands of raw job rows.
• Scheduled runs emit only material changes after the initial baseline.
• Technology-demand summaries reveal which infrastructure, security, observability, and data tools appear across matching roles.

Input

Provide up to 25 public ATS boards:

{
  "boards": [
    { "company": "GitLab", "provider": "greenhouse", "token": "gitlab" },
    { "company": "Lever Demo", "provider": "lever", "token": "leverdemo" },
    { "company": "Ashby", "provider": "ashby", "token": "Ashby" }
  ],
  "categories": ["sre", "devops", "platform", "cloud", "security", "ai_infrastructure", "data_engineering", "observability", "finops", "quant"],
  "onlyChanges": true,
  "emitBaseline": true,
  "stateKey": "my-watchlist"
}

The token is the final path component of a public board URL:

• https://boards.greenhouse.io/gitlab → gitlab
• https://jobs.lever.co/leverdemo → leverdemo
• https://jobs.ashbyhq.com/Ashby → Ashby

Arbitrary URLs are not accepted. Every upstream request is built from a fixed provider hostname and a validated token.

Output

Each dataset row is one company signal containing:

• current open and matching role counts;
• added, removed, and materially changed matching roles;
• an explainable 0–100 signal score and trend;
• matched categories, category deltas, technology-demand counts, and newly appearing or disappearing technology requirements;
• workplace type, remote status, employment type, and public compensation text when an ATS provides it;
• direct public source links and timestamps.

The OUTPUT record contains run health, source failure counts, budget status, and processing totals.

Unattended operation

The Actor is designed for scheduled, zero-touch operation:

• three bounded attempts for transient source failures;
• fixed public source hosts and hard response/job caps;
• previous state is preserved when a source fails;
• a fully failed run exits non-zero instead of publishing false removals;
• pay-per-event spending limits stop additional value delivery;
• state namespaces keep independent customer watchlists isolated;
• no API keys or private credentials are required from buyers.

The recommended schedule is once per day using the same stateKey. Downstream systems can consume the default dataset through Apify API, webhooks, Make, Zapier, n8n, or MCP.

Pricing setup

Recommended initial Apify Pay Per Event configuration:

• event name: company-scan
• event title: Company board successfully checked
• initial price: $0.05 per successful company check
• platform usage costs: pass through to the buyer
• permissions: limited
• Standby mode: disabled

At an 80% creator share, 5,000 successful company checks produce approximately $200 before tax when platform usage costs are passed through. Revenue is not guaranteed; usage and customer acquisition must be measured.

Local development

npm install
npm test
npm run check
npm run smoke

Run locally with an input stored at storage/key_value_stores/default/INPUT.json, or deploy with the Apify CLI:

apify login
apify push

Never commit an Apify token. Account ownership, KYC, tax information, payout setup, and acceptance of platform terms must remain with the human account owner.

Operational kill gate

Do not keep expanding the product without demand. After 30 days in the Store:

• continue if there are at least 3 external paying users or 500 paid company checks;
• otherwise keep the Actor stable, stop feature work, and test a different market hypothesis.