Product Hunt Launch Monitor — new launches by keyword

Watch Product Hunt's featured launches on a schedule and get only the NEW products matching your keywords since your last run — title, tagline, Product Hunt link and launch date.

A startup that launched this week is the warmest B2B lead there is: they have momentum, budget attention, and every tool decision still open. Agencies, dev-tool vendors, and service providers all know it — this Actor turns that into a deduplicated feed you can pipe into Slack, a Google Sheet, Make or Zapier.

What it does

1. Reads Product Hunt's public featured-launches feed (about 50 fresh products at any time).
2. Keeps only launches whose title or tagline matches your keywords (leave empty to monitor everything).
3. Remembers every launch it has already seen (per keyword set, in a persistent state store) and outputs only the NEW ones on every later run — schedule it daily and you get a clean drip of fresh launches, never duplicates.

Input example

{
  "keywords": ["ai agent", "analytics", "no-code"]
}

The defaults work as-is.

Output example

{
  "postId": "1188925",
  "title": "Astryx",
  "tagline": "A customizable, agent-ready open-source design system",
  "productUrl": "https://www.producthunt.com/products/astryx",
  "publishedAt": "2026-07-05T20:52:49-07:00",
  "matchedKeywords": ["agent"],
  "isNewSinceLastRun": true,
  "firstDetectedAt": "2026-07-06T16:00:00.000Z"
}

A SUMMARY record in the run's key-value store gives the counts: launches in feed, matching, new, total known.

How to use it

• Lead flow: schedule daily with your niche keywords → each morning you get only the launches that appeared since yesterday. Reach out while they're in launch mode.
• Competitive radar: put your category keywords in — you'll know the day a competitor launches.
• Trend pulse: run keyword-free and diff the volume of launches over time per monitorId.

Fair use & data

• Reads only Product Hunt's public launch feed — no login, no page scraping.
• No personal data: hunter/maker names are deliberately not collected; you get the product, not the person.
• The feed exposes currently featured launches (~50); the monitor's memory is what turns it into a complete over-time stream.

FAQ

Why keyword filtering instead of Product Hunt topics? The public feed is global; keywords match against title + tagline, which in practice is how you'd scan the homepage yourself. Multi-word keywords are matched as phrases.

First run floods me. That's the baseline (flagged isBaseline: true); disable includeBaselineInOutput to start silent.

How do I reset the memory? Set a new monitorId — each ID keeps its own baseline.