Hacker News Scraper: Stories, Comments & Mentions

Search every Hacker News story and comment since 2006, with points, comment counts, authors and links.

Built on the official Algolia-powered HN search API that YC publishes. No key, no login, and far better than scraping news.ycombinator.com, which is rate-limited and paginated 30 items at a time.

What you get

Field	Example
title, author	Show HN: I built…, pg
points, commentCount	412, 189
url, hnUrl	the linked article, and the HN thread
text	comment body or self-post text
createdAt	ISO 8601
isShowHN, isAskHN, tags
parentStoryId, parentStoryTitle	on comment rows, what thread it's from

Research vs monitoring

Research (default), relevance-ranked. "What has HN said about Postgres?"

Monitoring: sortByDate: true, newest first. "Did anyone mention us today?" Schedule it hourly and you have brand-mention alerts.

{
  "searchQueries": ["your-product", "your-competitor"],
  "sortByDate": true,
  "createdSince": "2026-07-01",
  "minPoints": 10
}

Front page without a query

Leave searchQueries empty and set contentTypes: ["front_page"] to list everything currently on the front page. Same for show_hn or ask_hn, a whole category, no query needed.

Filters

• contentTypes: story comment poll show_hn ask_hn front_page
• minPoints, minComments, applied by the API itself, so they cost you nothing
• createdSince / createdUntil

Who this is for

• Founders & marketers: every mention of your product or a competitor, as it happens
• Launch research: what got Show HN posts to the front page, and what didn't
• Trend research: how a technology's mentions changed over years
• Recruiters: the "Who is hiring?" threads, parsed
• AI / dataset builders: one of the highest-signal text corpora on the web

Every story and comment since 2006

Backed by the Algolia HN Search API, which is Hacker News' own official search backend, so this searches the full archive rather than the front page.

contentTypes picks what comes back: story, comment, poll, show_hn, ask_hn, front_page. ask_hn and show_hn are separate types on purpose, because "what has Show HN launched in this space" and "what has Ask HN asked about it" are different questions.

front_page is the one that works with no query at all: pass it alone and you get whatever is on the front page right now, which is a scheduled snapshot rather than a search.

Filters, all pushed down to the API so a narrow filter costs less rather than the same: minPoints, minComments, createdSince and createdUntil, plus sortByDate to switch from relevance to newest-first.

Per row: title, URL, author, points, comment count, creation date, the HN item URL, and the text of a comment or self-post. Comment rows carry parentStoryTitle, so a comment is never stranded without its context.

What it is good for. sortByDate with a brand name is a mention monitor, and it catches the thread while it is still on the front page. minPoints at a few hundred with a topic is a "what actually resonated" list rather than everything ever posted. Passing itemIds pulls specific threads whole.

No key, no login. The Algolia endpoint is public because it is what the HN site itself calls.

Notes

• Point and comment filters are pushed down to the API, so a narrow filter is cheaper, not slower.
• Comment rows carry parentStoryTitle, so you always know what a comment was replying to.
• HN scores keep moving after posting; a run captures the score at that moment.
• Set a max charge per run in the run options for a hard spend ceiling. The Actor honors it and stops cleanly.

Pricing

Pay per result. One row = one story, comment or item. Failed queries come back as error rows and are not charged.