Hacker News Scraper

Scrape Hacker News via the public Algolia HN Search API — no API key, no login, no anti-bot. Search stories, Show HN, Ask HN, comments, or the current front page; sort by relevance or date; filter by minimum points.

What it does

• Queries https://hn.algolia.com/api/v1/search (relevance) or search_by_date (newest first).
• Paginates automatically until it has collected maxItems.
• Returns clean, normalized JSON for each item (HTML stripped from story/comment text).
• Deduplicates by Hacker News objectID.

Input

Field	Type	Default	Description
query	string	"openai"	Keywords to search for. Empty = newest/front-page items for the chosen tag.
tags	select	story	story, show_hn, ask_hn, comment, or front_page.
sortBy	select	relevance	relevance or date (newest first).
minPoints	integer	—	Only items with at least this many points.
maxItems	integer	50	Max items to return (1–1000).
proxyConfiguration	proxy	off	Optional; the public API has no anti-bot, so no proxy is needed. Only enable it if you hit IP rate limits at very high volume.

front_page note: the front_page tag always returns the ~30 items currently on the HN front page (no keyword search). For topic searches use story/show_hn/ask_hn/comment instead.

Output

Each successful item:

{
  "ok": true,
  "objectId": "44159823",
  "type": "story",
  "title": "OpenAI ...",
  "url": "https://example.com/article",
  "author": "someuser",
  "points": 412,
  "numComments": 188,
  "createdAt": "2026-06-01T12:34:56.000Z",
  "text": "",
  "hnUrl": "https://news.ycombinator.com/item?id=44159823"
}

For comments, type is comment, title/url reflect the parent story, and text is the (HTML-stripped) comment body.

Some fields can be null depending on the item: url (Ask HN/text posts and self-posts have no external link), points and numComments (often absent on comments), and text (empty string for link stories). title, author, objectId and hnUrl are present for well-formed items but defensively default to null if the API omits them.

On failure or no results, a single diagnostic row is emitted with ok:false and an errorCode (e.g. NO_RESULTS, RATE_LIMITED, NETWORK) — and nothing is charged.

Delivery

Results always land in the run's dataset (JSON, CSV, Excel, XML via the API or the Console). Two optional inputs, notionConnector and notionParentId, will additionally write one Notion page per item when the run finishes — authorize a Notion MCP connector once under Settings → API & Integrations → MCP connectors, then pick it in the input form. Leave them empty to skip.

proxyConfiguration is off by default. The Algolia HN Search API is public and has no anti-bot, so a proxy adds latency and cost for nothing; only switch it on if you are running enough volume to hit IP-level rate limits.

What people use it for

• Brand and competitor monitoring — search story and comment for your product name and watch points / numComments to see which threads are actually getting traction.
• Launch tracking — show_hn sorted by date gives you every new Show HN as it lands.
• Hiring and demand signals — ask_hn surfaces "Who is hiring" and "Who wants to be hired" threads.
• Front-page snapshots — front_page returns the ~30 items on the HN front page right now, for a daily digest or a dashboard.
• Filtering the noise — minPoints drops everything below a score threshold before you are billed for it.

Pricing

$1.00 per 1,000 items ($0.001 each), with no run-start fee. Flat rate — no volume tiers, no plan gates — and you are charged only for items actually returned.

One charge per returned item (item event). Diagnostic and empty rows are never charged, so a query that matches nothing costs nothing. Because you pay per row, set maxItems to what you actually need — the default is 50 and the ceiling is 1,000 per run.