Job Intelligence: Greenhouse & Lever Jobs Scraper + AI Enrichment

Get every open job from any company that hires on Greenhouse or Lever, in one clean table, with AI-extracted seniority, skills, tech stack, salary, visa sponsorship and a 2-line summary. Turn on "Only new since last run" and schedule it to monitor hiring weekly.

No login, no browser, no proxies. Data comes from the companies' own public job-board APIs, the same source their careers pages use.

What you get

Field	Example
company, title, department	Anthropic · Staff+ Software Engineer, Research Systems · Engineering
location, remote_type	San Francisco, CA · remote / hybrid / onsite / unknown
posted_at, apply_url	2026-10-12T19:03:03Z · direct link
salary_min, salary_max, salary_currency, salary_period	320000 · 485000 · USD · year
salary_source	ats_field (from the board) or ai_from_text (stated in the description)
ai_seniority	intern · junior · mid · senior · staff · principal · lead · manager · director · executive
ai_role_family	engineering · data_ml · product · design · sales · marketing · …
ai_skills, ai_tech_stack	["dbt", "Airflow", "SQL"]
ai_visa_sponsorship	yes · no · not_mentioned
ai_years_experience_min	5
ai_summary	Builds billing data pipelines for usage-based pricing…
is_new	true if not returned by your previous runs
description	Full plain-text description (emails and phone numbers removed)

The AI never guesses salary: it only fills it in when a number is written in the posting.

Use cases

• Recruiters & talent teams: track who is hiring for which skills, at what pay, every week.
• Sales & GTM: hiring is a buying signal. "Company X opened 6 data-engineering roles using Snowflake" is a lead.
• Job seekers: filter by skills, seniority, remote or visa sponsorship across many companies at once.
• Market research: salary bands and tech-stack trends by role and location.
• Job boards & newsletters: a normalized, deduplicated feed with "new since last run".

Input example

{
  "companies": ["stripe", "https://boards.greenhouse.io/figma", "https://jobs.lever.co/palantir"],
  "keywords": ["machine learning", "data engineer"],
  "locations": ["London", "Remote"],
  "postedWithinDays": 14,
  "onlyNewSinceLastRun": true,
  "maxJobs": 500,
  "enableAI": true
}

Companies: paste a careers-page URL (boards.greenhouse.io/<company>, jobs.lever.co/<company>), or just a company slug like stripe; slugs are checked on both Greenhouse and Lever. Prefix with greenhouse: or lever: to force one.

Filters (keywords on title, locations, remoteOnly, postedWithinDays) are applied before AI, so you only pay for jobs you keep. Results are newest first, capped at maxJobs.

Pricing

Pay per event, no subscription:

Event	Price
Run start	$0.005
Job returned	$1.50 per 1,000
AI enrichment (optional)	$2.50 per 1,000

Examples: 1,000 jobs with AI ≈ $4.00. 1,000 jobs without AI ≈ $1.50. A weekly monitor of 20 companies returning ~100 new jobs ≈ $0.40 per run.

Have your own OpenRouter key? Add it in the input and AI enrichment is not charged; you pay OpenRouter directly. Set a maximum cost per run in Apify to cap spending.

Monitoring setup (recommended)

1. Create a task with your companies and filters, and turn on onlyNewSinceLastRun.
2. Add a weekly schedule.
3. Connect the output to Google Sheets, Slack or a webhook via Apify integrations, or pull it with the API.

FAQ

Is this legal / allowed? It reads public job-board APIs that companies publish so their postings can be displayed, at a polite rate (≤1 request/second per host). No logins and no personal data: emails and phone numbers are stripped from descriptions. You are responsible for how you use the data.

Which companies work? Any company whose careers page is hosted on Greenhouse or Lever. Many tech companies use one of them. Ashby support is planned.

Why did a company return nothing? It doesn't use Greenhouse or Lever, or uses a different slug. Open its careers page and copy the board URL into companies.

How accurate is the AI? It extracts only what's written in the posting, using a fast model at temperature 0. Fields it can't determine come back as unknown, not_mentioned or null rather than a guess. Spot-check before making important decisions.

LinkedIn or Indeed? Not supported, by design.

Support

Found a bug or need another ATS? Open an issue on the Issues tab with the company URL and run ID. Issues are answered within a few days.