Hacker News API

The Hacker News API reads the official public Algolia index of HN and returns stories as structured rows — title, outbound URL, points, author, comment count, timestamps, the HN discussion link and the index tags. Search every story ever posted, or take the current front page, in one call.

Give it a query and it searches the whole archive by relevance or by date; leave the query empty and it returns the live front page. You can target type explicitly to pull Ask HN or Show HN threads. Each run is one call returning up to 200 stories, read over the TLS tier — no browser, no scraping of the HTML site, just the JSON index HN's own search is built on. For mainstream coverage of the same stories, pair it with the Google News API.

$0.0003 per delivered story, up to 200 storys per run. Nothing delivered means nothing charged.

How the Hacker News API works

You optionally send a query, a type filter (story, Ask HN, Show HN, front_page), a sort (relevance or date) and a max_results up to 200. With a query it searches the archive; without one it returns the front page. Rows come back flat.

Two things to know:

Inputs

Every input is optional. query searches the archive; leave it empty for the front page. type targets a story kind (including Ask HN and Show HN), sort chooses relevance or date, and max_results caps delivery at 200 stories in a single call.

What one story looks like

Each story is one flat row: title, the outbound url (null for text-only posts), points, author, comments, an ISO created_at, the hn_url discussion link, any Ask-HN text with tags stripped, and the index tags. The HN item id is on every row so you can de-duplicate or join across runs.

What the Hacker News API costs

$0.0003 per delivered story ($0.30 per 1,000). Nothing delivered means nothing charged, and the $2 monthly free credit covers roughly 6,666 storys here. Volume tiers take up to 30% off.

$0.0003 per delivered story, $0.30 per 1,000. A full 200-story pull is six cents; polling the front page every minute for a day at 30 stories a time is roughly 43,000 rows — about

3. You pay per story returned, so a tight query that matches little costs little.

The $2 monthly allowance (no card) covers roughly 6,600 delivered stories, and volume tiers take up to 30% off. GET /v1/scraper/collectors returns your key's live price.

Hacker News API vs the official one

Being straight about it: this collector reads HN's official public search index, provided by Algolia, and that API is free and open. So is the Firebase HN API for raw items. If Hacker News is the only source you need and your volume is modest, calling either directly is a fine choice — there is no gatekeeping to route around here.

What this adds is shape and integration: typed rows against a published schema, search and front-page in one endpoint, and delivery under the same QuantumProxies key and per-row billing as your other collectors — with rotating exits so a heavy polling job is not throttled by a single IP. You are paying for one pipeline and one bill, not for access you could not otherwise get.

Versus calling the index yourself

HN's index is clean JSON, so the DIY cost is operational rather than parsing:

  • Rate limits at volume. Frequent polling from one IP eventually gets throttled; rotating exits keep a high-frequency front-page or search job flowing.
  • Two surfaces, one schema. Search results and front-page items differ slightly; folding them into a single flat row with a real created_at and a stripped text is boilerplate you would otherwise own.
  • One key for everything. If you already pull other collectors through QuantumProxies, adding HN here avoids a separate integration and lets you join it to the rest in one place.

If HN alone at low volume is the whole job, the free official index may be all you need; at scale or alongside other sources, this plus our residential proxies is the operational win.

What people build with it

Tech trend and topic monitoring

Search the archive for your product, competitor or topic and sort by date to catch new threads as they land, following hn_url straight into the discussion.

Front-page capture

Poll the front page on an interval and store each snapshot to reconstruct what led HN over a day, ranking by points and comments as they move.

Ask HN and Show HN mining

Target type to pull Ask HN questions or Show HN launches, using the text body and comment counts to study what the community asks and ships.

Limits, coverage and the legal bit

One call per run, up to 200 stories. It indexes stories and their metadata — points, comments, author, links, tags — not the full comment trees; a story's hn_url is where the thread lives. Front-page versus archive search is chosen by whether you pass a query. Runs parallelise up to your plan's rate limit — 60 a minute on pay-as-you-go, up to 1,200 on the top tier — which is also how you poll faster than one IP against the index could.

Legally: HN posts are public and reading public data is broadly lawful, but the content is user-generated, the index has its own terms, and rules vary by country. Background, not legal advice.

FAQ

Is the Hacker News API free?

$2 of usage a month with no card, which at $0.0003 per story is roughly 6,600 delivered stories. Hacker News' own Algolia search index is also free; this collector is worth it mainly when you want HN data as typed rows alongside other collectors, or polled harder than one IP allows.

Do I need an API key for Hacker News?

No. Your QuantumProxies API key is the only credential — the official HN index needs none either, and this collector reads it for you, so there is nothing extra to register.

Can I get Ask HN and Show HN?

Yes. Set the type filter to target Ask HN or Show HN specifically; an Ask HN row includes its text body with markup stripped, and every row carries the hn_url discussion link and comment count.

Does it return the comments?

It returns the comment count and a link to the discussion, not the full comment tree. The index is story-level; to read a thread you follow hn_url. That keeps a row flat and the per-story price low.

Related scrapers