Product Hunt CRM Lead Scorer

Turn messy Product Hunt launch rows into clean, deduplicated, scored startup leads that are ready for a CRM, spreadsheet, sales workflow, agency prospecting list, or VC/market research pipeline.

This Actor is not a Product Hunt page scraper. It is the cleanup and qualification layer after scraping or manual collection: one row per identifiable company, stable dedupe keys when rows include a Product Hunt URL, domain, slug, or product name, normalized contact/social evidence, transparent lead scoring, and flat export-friendly fields.

What it does

• Accepts Product Hunt-like launch rows in items mode.
• Normalizes common field variants such as name, title, product.name, website, websiteUrl, votes, votesCount, and makers.
• Deduplicates by canonical Product Hunt post slug first, then by normalized domain and product slug.
• Produces flat CRM-ready output fields.
• Adds deterministic leadScore, scoreReasons, icpFit, warnings, and an outreachHook.
• Keeps scoring transparent so you can audit why a lead was ranked high or low.

Who it is for

Use this Actor if you already have Product Hunt data from:

• another Apify Product Hunt scraper,
• a CSV/JSON export,
• manual research,
• an internal research workflow,
• or a no-code automation that collects launch rows.

Typical use cases:

• sales teams ranking newly launched SaaS/startup leads,
• agencies finding recently launched companies to contact,
• VCs and scouts triaging Product Hunt launches,
• growth teams preparing spreadsheet or CRM imports,
• builders cleaning output from raw scrapers before enrichment.

Input

Current supported input mode: items.

{
  "mode": "items",
  "maxItems": 500,
  "items": [
    {
      "name": "LaunchDeck",
      "productHuntUrl": "https://www.producthunt.com/posts/launchdeck",
      "websiteUrl": "https://launchdeck.io",
      "tagline": "AI follow-up engine for lean sales teams",
      "topics": ["Sales", "Artificial Intelligence", "CRM"],
      "upvotes": 420,
      "commentsCount": 37,
      "makerNames": ["Ada Lee", "Sam Ho"],
      "publicEmail": "hello@launchdeck.io",
      "emailSourceUrl": "https://launchdeck.io/contact"
    }
  ]
}

items must be an array of Product Hunt-like objects. The Actor is intentionally tolerant of common field names, so exact source shape can vary.

Output

Each output item is one normalized lead:

{
  "dedupeKey": "ph:launchdeck",
  "productName": "LaunchDeck",
  "productHuntUrl": "https://www.producthunt.com/posts/launchdeck",
  "productSlug": "launchdeck",
  "tagline": "AI follow-up engine for lean sales teams",
  "topics": ["Sales", "Artificial Intelligence", "CRM"],
  "upvotes": 420,
  "commentsCount": 37,
  "websiteUrl": "https://launchdeck.io/",
  "domain": "launchdeck.io",
  "makerNames": ["Ada Lee", "Sam Ho"],
  "publicEmail": "hello@launchdeck.io",
  "emailConfidence": "high",
  "leadScore": 90,
  "scoreReasons": ["domain:+20", "direct_email:+20", "maker:+10", "traction:+20", "icp_match:+20"],
  "icpFit": "high",
  "outreachHook": "LaunchDeck: AI follow-up engine for lean sales teams stood out with 420 Product Hunt upvotes.",
  "warnings": [],
  "evidenceUrls": [
    "https://www.producthunt.com/posts/launchdeck",
    "https://launchdeck.io/",
    "https://launchdeck.io/contact"
  ]
}

Output fields

• dedupeKey
• productName
• productHuntUrl
• productSlug
• launchDate
• tagline
• description
• topics
• upvotes
• commentsCount
• websiteUrl
• domain
• makerNames
• makerPhUrls
• xUrl
• linkedinCompanyUrl
• githubUrl
• discordUrl
• contactPageUrl
• publicEmail
• emailSourceUrl
• emailConfidence
• leadScore
• scoreReasons
• icpFit
• outreachHook
• warnings
• evidenceUrls
• scrapedAt

Scoring

Scoring is deterministic and explainable:

• domain:+20
• direct_email:+20 or contact_page:+12
• social:+10
• maker:+10
• traction from upvotes/comments up to +20
• ICP keyword/category match up to +20
• penalties for missing domain/contact/maker signals

The score is clamped to 0..100. Use scoreReasons and warnings to audit the result before importing leads into outreach tools.

Local development

From this Actor folder:

npm test
node src/main.js \
  --input ../../samples/product-hunt/product-hunt-crm-lead-scorer-input.json \
  --output ../../samples/product-hunt/product-hunt-crm-lead-scorer-output.json \
  --scraped-at 2026-07-05T00:00:00.000Z

Limitations

• This Actor does not scrape Product Hunt pages directly. Bring Product Hunt-like rows from another scraper/export/workflow.
• URL mode is not exposed yet; unsupported modes are rejected instead of silently returning empty output.
• Public or enriched emails are not guaranteed verified. Always review email evidence before outreach.
• The scoring model is deterministic and lightweight, not a machine-learning model.
• No external website enrichment is performed in this version.

Cost and permissions

• The Actor reads the provided Actor input and writes normalized rows to the default dataset; it does not request account-wide data access.
• It performs no browser crawling and no external HTTP scraping in the current version.
• Runtime is expected to be lightweight because it processes provided rows only.

Best results

For best results, include as many of these fields as possible in each input row:

• Product Hunt URL or slug
• product name
• website URL
• tagline or description
• topics/categories
• upvotes and comments count
• maker names or maker profile URLs
• public email or contact page URL
• social/profile URLs