Greenhouse Hiring Intent Signals — The SDR's Secret Weapon

This is NOT a generic job scraper. It's a high-speed, API-native extractor purpose-built to feed clean, actionable hiring signals directly into LLM prompt chains and CRM outreach workflows.

A 200-person Series B company posting 6 new "Sales Engineer" roles this week isn't just hiring — they just got budget approval for a tool your software replaces. That's the signal. This Actor surfaces it in under 60 seconds, with zero HTML bloat and zero wasted tokens.

---

Why Choose This Actor?

⏱ Time-Based Filtering (max_days_old)

Generic scrapers dump every job posted since 2019 into your pipeline — noise that burns API credits and wastes SDR time. This Actor's hard freshness gate drops any job last updated more than N days ago (default: 7). You only see signals from companies actively spending budget right now.

🏷 Auto Intent Categorization

Every job is classified into a high-level intent_category via zero-cost regex rules — no LLM call required:

Category	Example Signals	Your Play
Data & AI	Data Scientist, ML Engineer, AI Product Manager	Pitch your data infra / analytics tool
Revenue	Account Executive, SDR, VP of Sales	Pitch your CRM / sales engagement platform
Engineering	Backend Engineer, DevOps, Security Engineer	Pitch your developer tool / cloud service
Product	Product Manager, Program Manager	Pitch your project management / roadmapping tool
Marketing	Growth Marketer, Demand Gen, Content Strategist	Pitch your marketing automation / SEO tool
General	Everything else	Generic nurture

Route each category to a different email sequence automatically — no extra AI processing needed.

🌐 English-Only Guarantee

Greenhouse boards for global companies often contain duplicate job postings in French, German, Japanese, and other APAC/EMEA languages. These localised listings add zero value to your English-language outbound campaigns and burn expensive LLM tokens. This Actor auto-detects and filters out non-English job descriptions using a fast Unicode character-range heuristic, keeping your output lean and your token costs flat.

🧹 LLM-Ready Clean Text

Forget raw HTML with nested <div> tags, inline styles, and 5,000-character walls of text. This Actor strips all HTML via Cheerio and produces a clean job_summary_clean field — the first 400 characters of plain text, word-boundary truncated. Drop it directly into your GPT prompt:

"Write a cold email to the Head of Engineering at {company}, referencing this job opening: {job_summary_clean}"

---

Input Configuration

Parameter	Type	Required	Default	Description
company_boards	string[]	Yes	—	Greenhouse board tokens (the subdomain slug). e.g. "datadog" → https://boards.greenhouse.io/datadog
max_days_old	integer	No	7	Skip jobs last updated more than this many days ago. Set to 1 for today-only signals, 14 for a broader sweep. Max: 90.
target_departments	string[]	No	[] (all)	Case-insensitive department filter. Pass ["Engineering"] to only see engineering roles. Leave empty to scrape everything.

Example input:

{
  "company_boards": ["datadog", "stripe", "figma", "vercel"],
  "max_days_old": 3,
  "target_departments": ["Engineering", "Data Science"]
}

---

Sample Output

Here's exactly what you get — one flat, LLM-ready record per fresh signal:

{
  "board_name": "datadog",
  "job_title": "Senior Software Engineer - Data Platform",
  "department": "Engineering",
  "location": "New York, NY, United States",
  "updated_at": "2026-05-11T09:15:00-04:00",
  "days_active": 2,
  "intent_category": "Engineering",
  "job_summary_clean": "As a Senior Software Engineer on the Data Platform team, you will design and build the next generation of Datadog's petabyte-scale analytics infrastructure. You'll work closely with product teams to deliver real-time observability features used by thousands of enterprise customers. Strong experience with distributed systems, Kafka, and…",
  "apply_url": "https://boards.greenhouse.io/datadog/jobs/9876543"
}

Every field earns its place in your workflow:

Field	What it unlocks
days_active	Sort ascending. Hit 0–2 day signals first — those companies just opened budget.
intent_category	Route "Data & AI" signals to your data tool pitch. Route "Revenue" signals to your CRM pitch. Zero-runtime classification.
job_summary_clean	Drop into GPT-4o / Claude prompt. "Write a cold email referencing this opening…" No pre-processing needed.
apply_url	Cross-reference with your CRM. Existing customer? Deprioritize. Net-new logo? Gold.
department	Verify alignment with your ICP. You sell to Engineering leads? Filter target_departments: ["Engineering"].
location	Geo-qualify. Only selling in North America? Filter location downstream in Clay or Airtable.

---

How SDR Teams Deploy This

┌──────────────────┐     ┌──────────────────┐     ┌──────────────────────┐
│  Apify Scheduler  │────▶│  This Actor      │────▶│  Zapier / Make       │
│  (Daily 6 AM run) │     │  (Fresh signals) │     │  (Webhook trigger)   │
└──────────────────┘     └──────────────────┘     └──────────┬───────────┘
                                                             │
                                                  ┌──────────▼───────────┐
                                                  │  Clay / Airtable     │
                                                  │  (Enrich + ICP tag)  │
                                                  └──────────┬───────────┘
                                                             │
                                                  ┌──────────▼───────────┐
                                                  │  Smartlead / Outreach│
                                                  │  "{first_name}, saw  │
                                                  │   {company} is hiring│
                                                  │   a {job_title}..."  │
                                                  └──────────────────────┘

1. Schedule daily with max_days_old: 1 to catch every new posting.
2. Webhook to your enrichment layer — tag each company (ICP fit? Existing customer? Competitor?).
3. Auto-generate cold emails — pipe job_title + job_summary_clean + intent_category into your LLM prompt.
4. Prioritize by days_active ascending — freshest signals get first contact.

---

Local Development

git clone <repo-url> && cd greenhouse-intent-signals-scraper
npm install
npm start        # Run with tsx (no build required)
npm run build    # Compile TypeScript to dist/

When running locally without Apify input, the Actor auto-falls back to test boards (datadog, mailchimp) with max_days_old: 14 so you can verify end-to-end immediately.

---

Cost Efficiency

This Actor calls Greenhouse's public JSON API — one lightweight HTTP GET per company board. No headless browser. No Playwright. No expensive proxy rotation. Scraping 50 boards costs a fraction of what a single paginated website scrape would consume in Apify compute units.

---

Built for SDRs. Optimized for LLMs. No bloat. Just signal.