logo
EN

Hacker News Keyword Search

Find public Hacker News stories by keyword and publication period, ordered by relevance, date or points.

Developer
Pricing
$0.0005 / record
Last updated
2026-11-11

Overview

Collect public technology stories in a table that connects each topic to the stories discussing it. Titles, publication dates, article links and discussion activity help distinguish a widely discussed announcement from a result that merely mentions the same words.

The results bring topic discovery and community response together. They give researchers a starting set of reading material and give product teams context about which stories attracted attention, without requiring them to copy story details and engagement figures individually.

Data notes

Results come from a searchable index of public Hacker News submissions. Relevance ordering prioritizes search matches; chronological ordering compares publication times; hottest ordering compares points and then comment counts within a relevance-selected set of matching stories. Each row belongs to a keyword and a story. The same story can appear for different keywords, which preserves the relationship between a topic and its matches. Match ID stays stable for the same trimmed keyword and story; Story ID identifies the underlying story across topics.

Points and Comments reflect the values returned when collected and can change between runs. They describe accumulated discussion activity, rather than readership, product adoption or sentiment. Hottest results are not a complete popularity ranking of every matching submission and do not represent a time-decayed trending score. Published at refers to the submission's publication time in UTC, rather than the time an article was originally written. Calendar-date boundaries include the entire selected UTC days, while rolling periods are measured backward from the same run-start instant for every keyword. A missing count is represented by zero, and missing text by an empty value.

Article URL may be empty for a submission that consists of a question or discussion. Discussion URL still identifies its conversation. Story text contains the submission body when available, with formatting removed; it does not contain the linked article or individual comments. Search results are a selected view of matching submissions and do not establish that every relevant historical story has been included.

What the results look like

Each record represents one story matching one keyword; these sample activity figures can change over time.

KeywordStory IDTitlePointsCommentsPublished at
python45751400Uv is the best thing to happen to the Python ecosystem in a decade221413242025-10-29T18:57:29Z
python20915746Sunsetting Python 216167042019-09-09T06:43:24Z

Other columns include Article URL, Discussion URL, Author, Match ID and Story text.

Use cases

  • Prioritize reading: compare Title, Points and Comments to choose stories for closer review, then use Discussion URL to inspect the conversation.
  • Compare topics: group results by Keyword and compare the distribution of Points and Comments to understand the activity around the returned stories, while accounting for overlaps using Story ID.
  • Trace a topic over time: use Published at and Title to compare early discussions with later announcements; combine Story ID and Match ID to update repeated collections without duplicating the same topic-story pair.

Scope and boundaries

Public stories only, up to 20 keywords and 1,000 returned stories per keyword; hottest ranks at most 1,000 relevance-selected candidates, and search availability may return fewer matches.

  • Good for
  • Find stories about technologies, products or companies discussed by developers.
  • Collect comparable topic results with publication dates and discussion activity.
  • Compare recent coverage with earlier stories or examine topic activity within a publication period.
  • Do not use for
  • You need comment bodies, complete linked articles, or exhaustive historical coverage.
  • You need current front-page rank, a time-decayed trending score, or guaranteed highest scores across every matching story.

Failure handling

Failure and retry behavior declared by the developer. We recommend including it in your system prompt when integrating.

  1. 1For an empty result, try a broader or differently worded keyword; an empty result is valid.
  2. 2If search is temporarily unavailable or rate limited, retry later; earlier returned rows may remain after a failure.
  3. 3If the run times out, reduce the number of keywords or the per-keyword limit and retry.
  4. 4Repeated identical keywords are trimmed and processed once; use a list containing one string for a single search.
  5. 5Use valid YYYY-MM-DD dates with the start on or before the end; periods with no overlap return no stories.
  6. 6If oldest-first search is too broad, use a narrower publication period or more specific keywords.

Input

Parameters required to call this App, generated from the input.schema in manifest.json.

  • KeywordskeywordsRequiredarray<string>
    Search each string independently. Supply 1 to 20 non-empty strings; repeated identical keywords are processed once.
    Itemsup to 20 itemsExample["python"]
  • Limit per keywordlimit_per_keywordOptionalinteger
    Maximum stories returned for each keyword, from 1 to 1,000; defaults to 100. Searches with fewer available matches return fewer stories.
    Range1–1000Default100
  • Sort ordersort_byOptionalstring(enum)
    Choose how stories are ordered; defaults to relevance. Hottest sorts up to 1,000 relevance-selected matches by points, then comments, and does not guarantee the highest score across all matches.
    Allowed valuesMost relevantrelevanceNewest firstnewestOldest firstoldestHottest (points, then comments)hottestDefaultrelevance
  • Time rangetime_rangeOptionalstring(enum)
    Restrict publication time to a rolling period ending at run start; defaults to all time. If custom dates are also supplied, only the overlapping period is searched.
    Allowed valuesAll timeallPast 24 hourslast_24hPast 7 dayslast_7dPast 30 dayslast_30dPast 365 dayslast_yearDefaultall
  • Custom date rangedate_rangeOptionalobject
    Optionally restrict publication time by UTC calendar dates, including both boundary days. Missing dates leave that side open; combined with Time range, only the overlap is searched.

Output

Field structure of a single record, generated from the output.schema in manifest.json.

  • Match IDidstring
    Stable identifier for the exact trimmed keyword and story pair.
    Example759cbfc402d5206e48b00a01788e8f5af14e87a83916eb4f340586d316485a0e
  • Keywordkeywordstring
    The trimmed search string that matched this story.
    Examplepython
  • Story IDstory_idstring
    Hacker News story identifier, shared across matching keywords.
    Example45751400
  • Titletitlestring
    Story title; empty when unavailable.
    ExampleUv is the best thing to happen to the Python ecosystem in a decade

Sample data

Generated from output.schema. Fields without a sample value use structural placeholders.

output.example
{
  "id": "759cbfc402d5206e48b00a01788e8f5af14e87a83916eb4f340586d316485a0e",
  "keyword": "python",
  "story_id": "45751400",
  "title": "Uv is the best thing to happen to the Python ecosystem in a decade",
  "url": "https://emily.space/posts/251023-uv",
  "hn_url": "https://news.ycombinator.com/item?id=45751400",
  "author": "todsacerdoti",
  "points": 2214,
  "num_comments": 1324,
  "created_at": "2025-10-29T18:57:29Z",
  "story_text": "<story_text>"
}

Integration

This App can be integrated via MCP, API, SDK, or file export. All channels share the same capabilities and pricing. Every request authenticates with the Authorization: Bearer header using an API key (long-lived, created in the Data Hub Console); MCP clients can also sign in with OAuth, no key required. More options such as CLI and Skill are on the way. This App runs asynchronously and can be canceled midway: After you cancel, the run stops. Records already delivered are kept and billed.

With the MCP (Model Context Protocol), you can call this App directly from AI clients like Claude and Cursor. Pick your client and auth mode, then copy the config below.

Client config

Replace the value after Bearer with your long-lived API key. Works in any client, CI, or headless environment.

mcpServers config
{
  "mcpServers": {
    "octoparse_datahub": {
      "type": "http",
      "url": "https://mcp-v2.octoparse.com?pin=donglichun/hacker-news-keyword-search",
      "headers": { "Authorization": "Bearer <YOUR_API_KEY>" }
    }
  }
}

Let AI set it up for you

Don't want to edit configs by hand? Copy the install prompt and paste it into any AI client. It will complete the setup its own way. (By default it guides you through signing in, and never asks you to send your API key in the chat.)

detail.access.mcp.composeHint

Pricing

Per record returned

Charged by the number of records successfully returned. Failed runs are not charged.

$0.0005/ record

Multiple billing events accumulate independently. See each item for details. Failed runs are not charged.

View credits

Try it now

Fill in the parameters and run. Results come from a real call.

Example
Parameters are validated against input.schema before submission
$0.0005 / record