Hacker News Trends Exporter

Export current public Hacker News stories, scores and discussion links into a tidy JSON dataset and UTF-8 CSV. The Actor uses the official Hacker News Firebase API, with no account, AI model or secret token. It is designed for transparent topic research, not for predicting popularity or sales.

What it does

Choose one of six official feeds: top, new, best, ask, show or jobs. The exporter fetches only the first bounded set of public story IDs and their corresponding public metadata. It keeps story titles, original links, Hacker News discussion links, points, comment counts and publication times.

Filter by literal words in the story title, minimum points, minimum comments and a maximum age in days. You can scan up to 60 IDs and save up to 50 matching stories, never more than you scanned. The original feed order is preserved; the source may change between runs.

Input

The default input uses the top feed, scans 25 story IDs and saves up to 20 results. The number of scanned IDs controls API requests and run time. An age window of zero means no age filter. Job postings often have no point or comment count, so the defaults do not exclude them.

The Actor accepts only these fixed feed names and numeric limits. It does not accept a URL or hostname to fetch, and it does not search private comments, user accounts or personal information.

Output

Dataset rows contain ID, title, item type, public score, comment count, publication timestamp, age in hours, source domain, original URL when it is a safe HTTP(S) address, and the Hacker News discussion URL. It never visits links included in stories.

The default key-value store contains STORIES_CSV, REPORT and EXPORT_SUMMARY. The CSV is spreadsheet-safe: quoted values and escaped formula prefixes. The summary records the number scanned, the number matched and saved, and whether the public feed scan was partial. It also includes a small breakdown by linked domains.

Pricing and cost control

Proposed Pay-per-event price: $0.0003 per saved story. Only saved output rows are intended to be charged to a paying user; excluded records and unsuccessful source reads do not produce an item event. This is a pricing candidate, not proof of an active Apify Store price or profitability.

The runtime caps requests at 61 logical reads, output at 50 stories, and uses a fixed public API. No paid proxy, API subscription or model inference is required by the code. Actual platform charges and net payout must be measured in a verified cloud run before publication.

Limitations and fair use

Hacker News rankings, scores and comment counts change frequently. The tool checks only the current public IDs in the selected official feed; it does not search historical posts beyond that feed or perform AI classification. An empty filtered dataset is a valid result, not a demand forecast. Never present test runs as paying users or revenue.

Source: https://github.com/HackerNews/API . Respect provider availability and rate limits; a provider error aborts the run rather than returning a partial paid dataset.

Example

Use examples/input.json to request a small top-stories batch. The dataset and CSV contain actual public story metadata returned at run time, while examples/output.json contains three public source examples mapped directly from the official API. This is an illustration of the output shape, not evidence that the Actor handler or Apify SDK has executed successfully.