Best Hacker News APIs & scrapers in 2026

People search for a Hacker News API when they want Show HN launches, Ask HN threads, Who is Hiring rows, or comment trees in a product.

Usually they need one of three things: wire HN into a request handler, run overnight batch dumps, or assemble HN's public endpoints themselves. Those are different products. Mixing them up gets you id lists without titles, comment trees walked node by node, search hits that don't match your story schema, and favorites that never appear in the JSON APIs.

This is a coverage and ergonomics comparison as of October 2026. Re-check each vendor's docs and pricing before you commit budget.

Social Fetch is listed first because it is our product, and because it suits sync REST over HN search, ranked feeds, stories, comment pages, user history, public favorites, and structured Who is Hiring rows, especially if you already normalize Reddit, X, or Web under the same key. HN's own public read APIs suit teams where HN is the only network and who want to own every hop. Apify suits a custom crawl field nobody documents. Raw browser stacks suit teams that need HTML or screenshots rather than a typed story object.

The short list

1. Social Fetch: sync REST for search, feeds, stories, comments, users, favorites, and hiring.

2. HN public read APIs: assemble item, user, feed, and search clients yourself.

3. Apify actors: scheduled dumps and niche field extraction.

4. RapidAPI listings: a marketplace for weekend discovery.

5. Proxy and browser stacks: raw page capture when you will parse the result yourself.

Full route list: Hacker News platform page. Hands-on walkthrough: How to scrape Hacker News data.

What teams actually buy an HN API for

Ignore "we support Hacker News" marketing grids. The jobs that show up in production tickets look like the list below.

• Launch or brand mention watch: search hits for a product name, then story cards and comment pages for the ids you keep.

• Morning digest: top, new, show, ask, and jobs rows with title, score, URL, and comment count rather than bare ids.

• Sentiment on a thread: cursor-paginated comments for a known story id, without timing out on busy trees.

• Author research: a user card, then submissions or comments for prolific posters.

• Hiring intel: monthly Who is Hiring comments as rows you can filter, not one giant text blob.

• Saved-signal mining: public favorites for a username, with private favorites as a typed miss.

If a provider cannot clear those jobs, you will rebuild the missing pieces yourself, which is how free official access turns into a second product.

How I rank Hacker News data APIs

I apply the same evaluation criteria to every entry. Each entry below carries the same labeled fields: Best for, Strengths, Limitations, and Pricing.

1. Assembled objects vs plumbing. Do feeds return hydrated stories, or id lists you must fan out?

2. Search and story shape. Can one parser handle hits and story gets, or do you maintain two models?

3. Comment pagination. Busy threads need pages and cursors, not a single recursive dump that blows timeouts.

4. Hiring and favorites. These are the surfaces DIY work most often skips, and they are also high-signal.

5. Typed misses. lookupStatus and empty hits beat silent HTML.

6. Sync vs async. Product handlers prefer one HTTP round trip, while overnight dumps can live on actors.

7. Multi-platform tax. If Reddit already shares your envelope, ask whether HN forces a second client.

8. Compliance fit. This is public HN data under Y Combinator's terms, and you own how you use it. I am not your lawyer.

1. Social Fetch

Best for: product teams and agents that need search, digests, comment pages, author history, public favorites, and Who is Hiring beside other networks, without owning HN glue code.

Strengths: Social Fetch exposes documented /v1/hackernews/** routes with the shared { data, meta } envelope, including meta.requestId, meta.creditsCharged, and typed outcomes where routes expose lookupStatus. Authenticate with x-api-key. One key covers 23 platforms and 232 documented operations, there is an MCP server for AI agents, and new accounts get 100 free credits with no card. Failed lookups are free, and pay-as-you-go credits never expire. In practice that buys you the following.

• Search with type, author, domain, URL, and minPoints filters for brand and launch monitors.

• Ranked feeds (top, new, best, ask, show, jobs) with hydrated items, not id-only lists.

• Story and comment pages for discussion jobs, plus a polymorphic item get when the id type is unknown.

• User card with submissions, comments, and public favorites for author analysis.

• Who is Hiring as structured rows with best-effort company, role, and location fields, and the full plain-text body as source of truth.

Limitations: Social Fetch covers 23 platforms, not forty, and coverage is public data only. Private favorites return lookupStatus: private_or_unavailable, and upvoted pages are not supported. Each successful request or cursor page is credit-metered, so confirm meta.creditsCharged.

Pricing: 100 free credits with no card, then from $1.35 per 1,000 credits on the Scale annual plan, or $1.64 per 1,000 on the Scale pack, as of October 2026 on pricing.

Hub: Hacker News platform page. Guide: How to scrape Hacker News data.

2. HN public read APIs

Best for: HN-only scripts where you want first-party public JSON and will own assembly, retries, and schema glue.

Strengths: Hacker News publishes public read surfaces for items, users, and ranked feed id lists, plus a separate public search surface. That stack is genuinely useful, and it suits internal scripts where HN is the only network, volume is modest, and a staff engineer already knows the quirks.

Limitations: it is incomplete as a product API. Feed endpoints give you ids that you hydrate item by item, comment threads are trees you walk, and search hits rarely match your story object one to one. Public favorites live on HTML pages, not the JSON paths used for items, and Who is Hiring is a monthly comment thread you have to discover and parse. It loses when a customer-facing feature needs one error shape, one credit meter, and the same parser you use for Reddit next quarter.

Pricing: free to call, with your own infrastructure and engineering time as the cost.

3. Apify actors

Best for: scheduled dumps, niche HTML fields, and scrapers you want to configure as Actors rather than maintain as a fleet of browsers.

Strengths: Apify hosts HN scrapers that start a run, poll, and download a dataset. Marketplace breadth helps when no managed REST route exposes the field you need.

Limitations: schemas differ by actor author, and sync product UIs that need a single HTTP round trip fit poorly. See Social Fetch vs Apify.

Pricing: billing mixes a platform subscription, compute, and per-event fees, and monthly plan credits expire. Model the listing you will actually run.

4. RapidAPI listings

Best for: finding a niche HN endpoint over a weekend without signing an enterprise contract.

Strengths: RapidAPI-style marketplaces aggregate third-party HN scrapers behind one billing UI, which is fine for disposable prototypes.

Limitations: there is no shared schema across publishers, and pagination, freshness, and empty-result behavior vary by listing. Adding Reddit or Truth Social beside HN means more listings and more adapters. Longer take: Social Fetch vs RapidAPI.

Pricing: set per listing by each publisher.

5. Proxy and browser stacks

Best for: raw HTML, screenshots, or one-off captures you will parse yourself.

Strengths: proxy and browser products are infrastructure, so you control exactly what gets fetched and can archive pages as captured.

Limitations: they do not give you a Who is Hiring row with lookupStatus, so you own the parser and the field drift. Buy them when you need capture, and buy a data API when you need a query returned as JSON.

Pricing: usage-based and set by each provider.

DIY public HN APIs vs a managed product API

A few vendors also sell HN-only REST catalogs covering search, lists, items, and users. They can be fine if HN is the entire product, but they leave you with a second vendor the day you add Reddit brand mentions or Web extraction, and many skip favorites and structured hiring. Compare field coverage carefully. The practical differences between HN's public APIs and Social Fetch fall along these lines.

• Hydrated front-page digest: with the public APIs you fan out id to item yourself, while Social Fetch feeds return items.

• Keyword search plus story schema: the public APIs need a separate search client and a mapping, while Social Fetch uses one envelope.

• Busy comment threads: the public APIs mean recursive walks or custom paging, while Social Fetch offers cursor pages.

• Public favorites: the public APIs mean an HTML scrape, while Social Fetch has a dedicated route with a typed private miss.

• Who is Hiring rows: the public APIs mean discovering the thread and parsing comments, while Social Fetch has a structured route.

• Same key as Reddit or Web: not available with the public APIs, and available with Social Fetch.

• Support requestId: you invent it on the public APIs, and Social Fetch returns meta.requestId.

Which option fits which job

• Brand or launch mention watch: Social Fetch search.

• Morning top or show digest: Social Fetch feeds.

• Comment sentiment on a known story: Social Fetch story comments.

• Author analysis: Social Fetch users and activity routes.

• Monthly hiring parse: Social Fetch who-is-hiring.

• Overnight HTML archive: Apify or a browser stack.

• HN-only personal script: the public HN APIs.

Questions before you buy

1. Do you need assembled stories, or are bare ids enough? Digests need hydration.

2. Is search part of the job? Many DIY stacks under-build search filters.

3. Will HN sit beside Reddit or Web? Prefer one schema if yes.

4. Sync or batch? Product handlers prefer sync.

5. Do credits expire? Spiky research sprints hate monthly resets.

6. Do you need hiring or favorites? Ask vendors explicitly, because those are the common gaps.

Frequently asked questions