Product Hunt Leaderboards Scraper — Daily, Weekly & Monthly Launch Rankings

Scrape Product Hunt leaderboards: the daily, weekly and monthly rankings of launches, with each launch's rank, votes (as displayed), comments, topics and link. Turn on launch details for makers, hunter, description, pricing and product stats.

Use dated exports to track a launch's ranking over time, compare weeks or months, or analyze which topics and makers show up at the top. One record is delivered per launch and per leaderboard: the same launch appearing in the daily and the weekly leaderboard gives two records, each with its own rank. Votes and comments move in real time on Product Hunt; every record carries a scrapedAt timestamp: the time the leaderboard page that served this record was read (each page of a long leaderboard has its own time). With details on, detailsScrapedAt is the time that launch's detail query answered.

Try it with 20 launch records

1. Paste a leaderboard URL into Leaderboard URLs, e.g. https://www.producthunt.com/leaderboard/weekly/2026/43.
2. Leave Maximum number of launch records at 20.
3. Start the Actor and open its dataset.
4. Read Run summary before exporting: a run can deliver results and still contain partially collected leaderboards (see Coverage below).

{
  "startUrls": [{"url": "https://www.producthunt.com/leaderboard/weekly/2026/43"}],
  "includeDetails": false,
  "maxItems": 20
}

The form prefill is 20 records; the schema default is 500. Set maxItems explicitly in API calls; 0 removes the limit. You are charged per delivered record, plus a $0.005 start fee per run.

Choose leaderboard URLs

Input	Data	Limit
Daily /leaderboard/daily/YYYY/M/D	The full day's leaderboard: first page + pagination until the site says it is done	A date still in progress returns only what exists so far
Weekly /leaderboard/weekly/YYYY/W	The full week's leaderboard, paginated the same way	Rank gaps left by the site are reported (rankGaps), never silently filled
Monthly /leaderboard/monthly/YYYY/M	The month's leaderboard, paginated the same way	Same, plus duplicate ranks reported (duplicateRanks)
/all suffix or any other page	—	Rejected with a reason; it never blocks the other URLs

includeDetails costs one extra request per delivered launch. reviewsCount is always null: the number of reviews displayed on a product page is not available without loading each 1.3 MB product page, and the launch query returns a different number (reviewersCount) that would be wrong under this name (measured: CleanShot displays 55 reviews, the query returns 48). The exact rating (rating) and follower count (followers) are delivered when details are on.

Output

Fields	Meaning
launchId, launchSlug, productId, productSlug	Launch and product identifiers. A product can have several launches; slugs can differ (launch notra-5 belongs to product notra-3).
name, tagline, url	Launch name and tagline as displayed (trimmed), and the launch page URL.
leaderboard, leaderboardDate, rank	Which leaderboard the record comes from (first day of the period in leaderboardDate) and the launch's rank in it — the site's rank, never recomputed.
dailyRank, weeklyRank, monthlyRank	Each period's rank as carried by the data, when present.
votes, commentsCount	As displayed on the leaderboard: votes is the current score (latestScore), never the launch-day score.
topics	All topics from the site data, including contest topics the leaderboard hides from display.
featuredAt, thumbnailUrl, scrapedAt	Launch time exactly as the site gives it (Pacific time with offset), thumbnail, and the UTC time the leaderboard page that served this record was read.
detailsScrapedAt	With includeDetails, the UTC time the launch detail query for this record answered (present only when detailsSource is postPage).
detailsSource	postPage when the detail fields were fetched, leaderboard when they are null (details off or failed).
description, pricingType, makers, hunter	Launch details (with includeDetails). makers are the launch's makers, not the product page's author list; hunter is null when the site gives no hunter.
websiteUrl, productName, productTagline, followers, rating, productLaunchCount	Product data (with includeDetails). websiteUrl is the product's website, which can differ from the link of this particular launch (Inferock: launch link inferock.ai, product website on GitHub). followers is the exact count (the page rounds, e.g. 3564 → "3.6K"); rating is null when the product has no reviews; reviewsCount is always null (see above).

Real output example

Collected on 28 October 2026 — demonstrates the format, not today's scores.

{
  "launchId": "1265035",
  "launchSlug": "inferock",
  "name": "Inferock",
  "tagline": "Serverless inference, with credits back for failed calls.",
  "url": "https://www.producthunt.com/products/inferock?launch=inferock",
  "productId": "1291234",
  "productSlug": "inferock",
  "leaderboard": "daily",
  "leaderboardDate": "2026-10-14",
  "rank": 2,
  "dailyRank": 2,
  "weeklyRank": 3,
  "monthlyRank": 51,
  "votes": 262,
  "commentsCount": 26,
  "topics": ["Developer Tools", "Artificial Intelligence", "Google Cloud Run"],
  "featuredAt": "2026-10-14T00:01:00-07:00",
  "thumbnailUrl": "https://ph-files.imgix.net/7a48325f-d5b7-4547-9aaf-aae3b709d967.png",
  "scrapedAt": "2026-10-28T07:51:19.774110+00:00",
  "detailsSource": "postPage",
  "description": "Get free signup credits to use all available models on Inferock.\n\nBring your own keys: run your OpenAI, Claude or Gemini accounts, see where calls fail and what they cost, and get credits back on eligible failed calls.\n\nServerless: pay-per-token LLM inference. Get signup credits to start using.",
  "pricingType": "free_options",
  "makers": [
    {"name": "Hamza Afzal Butt", "username": "hamza_afzal_butt",
     "headline": "B2B SaaS Growth Marketing Consultant",
     "profileUrl": "https://www.producthunt.com/@hamza_afzal_butt"},
    {"name": "Sumedh Bengale", "username": "sumedh_bengale",
     "headline": "Dev Rel Assoicate at Inferock",
     "profileUrl": "https://www.producthunt.com/@sumedh_bengale"},
    {"name": "Bharath Koneti", "username": "bharath_koneti",
     "headline": "Co-founder Inferock AI",
     "profileUrl": "https://www.producthunt.com/@bharath_koneti"}
  ],
  "hunter": {"name": "fmerian", "username": "fmerian",
             "profileUrl": "https://www.producthunt.com/@fmerian"},
  "websiteUrl": "https://github.com/inferock/inferock-bench",
  "productName": "Inferock",
  "productTagline": "An independent receipt for every LLM API call",
  "followers": 874,
  "reviewsCount": null,
  "rating": 4.0,
  "productLaunchCount": 2
}

This is the record as collected, except that the makers list is shortened here (makers had 4 entries for this launch). The "Google Cloud Run" topic is real data the leaderboard page hides from display — this Actor keeps it.

Access and Cloudflare challenges

Product Hunt filters server IP addresses with Cloudflare challenges ("Just a moment…"). The Actor runs through the Apify proxy by default (datacenter): when a request is challenged, it does not treat the response as data or as an empty page — it opens a new address and retries the same request, and after three consecutive challenges on datacenter addresses it switches to residential IPs for the rest of the run. Transient network failures through the proxy (SSL resets, dropped connections, 502/503/504 gateways without a challenge page) get the same treatment: new address, same request. Rate limiting (HTTP 429) and a 200 answer that is not readable JSON are retried the same way. Every request gets at most 8 attempts in total; a still-challenged URL is reported as blocked (blocked_by_site) and a still-failing one as network_error, each with the count of attempts and a dedicated warning — the run says so instead of pretending success. If you select your own proxy group or your own proxies, your choice is kept as-is for the whole run.

This address rotation can slow a run down (every request keeps a 2-second spacing). Once maxItems is reached, no further leaderboard page is fetched at all. The extra proxy cost is included in the per-record price. The run output reports requestsByAccess (requests made through datacenter / residential / none, or custom for your own proxies or another proxy group you chose) and challenges (how many Cloudflare challenges were answered by a new address).

Coverage and run summary

Product Hunt does not announce a total, so completeness is measured from what the site actually serves. A leaderboard is marked done when its pagination ends normally (hasNextPage=false, no failed page, no stale query id, no looping cursor); anything else is partial_collection with the reason.

Product Hunt itself skips ranks, e.g. there is no #10 on 14 Oct 2026, and September 2026 has more than 20 gaps. Missing ranks are not a partial collection: they are listed per URL in the run output (rankGaps) so you can account for them. The same applies to duplicate ranks — the site sometimes serves two launches at the same rank (twice at #80 in week 43); both are delivered and the rank is listed in duplicateRanks. Ads inserted in the feed (adsSkipped) are counted and never delivered. A future date or an empty period is no_products with a reason; if every URL comes back empty the run fails with a clear message instead of a silent success.

If Run summary stays blank, open the run's Storage → Key-value store tab and click the eye icon on the SUMMARY row. Check per-URL statuses before treating a dataset as complete.

Known limits

• reviewsCount is always null (explained above); rating and followers are exact.
• Product Hunt itself skips and repeats ranks (see Coverage); the delivered records follow the site's data, gaps and duplicates are reported per URL.
• Product Hunt changes internal query ids when it updates the site. If that happens, requests fail with a dedicated warning ("Product Hunt changed an internal query id; this Actor needs an update"), leaderboards already fetched are still delivered, and the affected URLs are reported — never a silent empty success. If the launch detail query is the one affected, details are turned off for the rest of the run at the first failure (one wasted request, not one per launch): launches keep coming with detailsSource: "leaderboard", and the run summary says so.
• Requests are spaced 2 seconds apart; a closed day takes 1 to 3 requests, a full week about 10, a month several dozen — plus one request per launch when details are on. Cloudflare challenges add retries and address rotations (see Access).

Pricing

$1.50 per 1,000 delivered launch records plus a $0.005 start fee per run: $0.035 for 20, $1.505 for 1,000. The start fee is charged once per GB of run memory (the default is 1 GB, so once per run). Apify subscribers pay less automatically, on both events: 10% off on the Bronze plan, 20% on Silver, 30% on Gold and above. Standard Apify platform usage, including the default proxy and any residential escalation, is included in this price. A run delivering no records only pays the start fee; each delivered record (one launch in one leaderboard) is one chargeable event.

If you set a maximum cost per run, the Actor stops cleanly when it is reached: the run succeeds, the records already delivered stay in the dataset, the affected leaderboard URLs are marked stopped_at_limit with how many of the launches read were delivered, and the status message says the maximum charge per run was reached.

Personal data

With includeDetails, records include the public profile data Product Hunt shows for a launch's makers and hunter: names, usernames, headlines and profile links. This data is public on Product Hunt, but it is still personal data. You are responsible for how you use it under the GDPR and the other laws that apply to you: have a legitimate purpose, keep only what you need, and do not use it for unsolicited outreach or marketing. Leave includeDetails off if you do not need it.

Automation and support

Use the Apify API to run the Actor on a schedule (e.g. every day for the previous day's daily leaderboard) and keep dated exports to follow rankings over time.

For a reproducible problem, open an issue with the leaderboard URL, run ID and expected result. Do not include API tokens. Site changes can affect collection; check the run summary before using the results.

Français

Collectez les classements Product Hunt (jour, semaine, mois) : rang, votes affichés, commentaires, thèmes et, en option, les détails de chaque lancement (makers, chasseur, description, statistiques produit). Une ligne par lancement et par classement. Les classements sont paginés jusqu'au bout, dans la limite de maxItems : 20 dans le formulaire, 500 par défaut dans l'API (un mois dépasse 600 lancements), 0 = sans limite ; les rangs sautés par le site sont listés (rankGaps), les rangs doublés livrés deux fois (duplicateRanks). Le proxy Apify est activé par défaut : le site filtre les serveurs par défis Cloudflare, l'actor change d'adresse et passe aux IP résidentielles au besoin (coût inclus). reviewsCount est toujours null (le nombre affiché n'est pas disponible sans charger chaque fiche produit). scrapedAt est l'heure de lecture de la page de classement qui a fourni la ligne, detailsScrapedAt celle de la fiche du lancement. Les détails contiennent des données personnelles publiques (noms, pseudos, titres et profils des makers et du chasseur) : vous êtes responsable de leur usage (RGPD, lois locales) ; ne les utilisez pas pour du démarchage non sollicité.

Product Hunt is a trademark of its owner. This Actor is not affiliated with, endorsed by or sponsored by Product Hunt.