ShortMax Scraper — Catalog, Genres, Views, Likes & Episodes

Scrape the ShortMax vertical short-drama catalog (shortmax.com → shorttv.live) without the app: titles, genres and categories, view & like counts, episode counts, synopses, creators and covers — structured and ready for analysis. Reads the site's own server-rendered HTML, no browser, so runs are fast and cheap. Export as JSON, CSV, Excel or XML, or pull straight through the Apify API into your app, agent, notebook or spreadsheet.

ShortMax is one of the big three Western-facing short-drama apps alongside ReelShort and DramaBox, and until now had no structured-data source at all. Output is normalized to the same per-title shape as the ReelShort and DramaBox scrapers, so the three merge into one cross-platform leaderboard.

No per-datapoint games, never charged for failures. Flat $2 per 1,000 titles ($0.002 each), billed only per title actually stored. Empty pages, failed requests and titles with no metrics are never stored and never charged.

What it scrapes

Field	Description
id / title / url	Numeric ShortMax title id, name, canonical URL
views / follows	View and like counts (see the honesty notes below)
viewsRaw / followsRaw	The site's own abbreviated strings ("88.1K"), so you can audit the precision loss
episodes	Episode count (Full mode)
genres	ShortMax's genre taxonomy — 56 site-wide, typically 1–2 per title
tags	ShortMax's second, separate category taxonomy — 53 site-wide, exactly one per title (Full mode)
description	Full synopsis
creator	The single writer/creator credit ShortMax publishes
cover	Cover image URL, resize query string stripped
releaseDate / releaseDateSource	Approximate — see below
section / rank	Homepage rail and position (Trending mode)
language / partial / scrapedAt	Locale, whether the record is listing-only, capture time

Modes

• Full catalog — walks every page of /dramas. The English catalog is 505 titles; other locales differ (see Languages).
• By category / By genre — ShortMax has two independent taxonomies and this Actor exposes both. Pass a slug (modern), a slug-with-id (modern-200052), a bare id (200052) or the display name (Modern) — every value is resolved live against the site, and an unrecognized one is logged alongside the complete list of what is available.
• Trending — the 9 curated homepage rails (New Release 🆕, Most Popular 🔥, War God ⚔️, …), the only editorial ordering ShortMax publishes. Every record carries its rail name and 1-based position.
• Specific title URLs — scrape exactly the /drama/... URLs you provide.

Detail level. Full visits each title's own page — the only source of episode counts and the category breadcrumb (1 request per title). Listing only reads views, likes and synopses straight off genre list pages, 10 titles per request, for a much cheaper run without episode counts or categories. Listing only applies to genre-based collection: catalog and category pages are plain image-and-title grids with no metrics on them at all, so the Actor falls back to Full for the other modes rather than billing you for metric-less rows.

Languages

ShortMax runs 19 locales, and each one is a different library, not a translation — the Spanish and Japanese catalogs contain entirely different titles from the English one. Approximate sizes: English 505, Traditional Chinese ~900, Spanish ~400, Japanese ~300. Hindi is configured on the site but its catalog is empty; the Actor reports zero results for it rather than failing.

Input example

{
    "mode": "byGenre",
    "genreIds": ["werewolf-200071", "mafia-200080"],
    "language": "en",
    "detailLevel": "full",
    "maxItems": 200
}

Output example

{
    "platform": "shortmax",
    "id": "32605",
    "title": "SSS-Rank: The Slum-Born Thunder God",
    "url": "https://www.shorttv.live/drama/sss-rank-the-slum-born-thunder-god-32605",
    "language": "en",
    "cover": "https://akamai-static.shorttv.live/images/cover/2026/08/18/f07994a8104e45818b452123087d0913.jpg",
    "description": "In a world where Awakening rank decides everything, Nate Ryder is dismissed as a worthless failure…",
    "genres": ["Modern", "Reversal of Fortune"],
    "tags": ["Urban Fantasy"],
    "views": 88500,
    "follows": 29500,
    "viewsRaw": "88.5K",
    "followsRaw": "29.5K",
    "episodes": 61,
    "creator": "Grace Whitman",
    "cast": [],
    "releaseDate": "2026-08-18",
    "releaseDateSource": "cover-asset-path",
    "section": null,
    "rank": null,
    "partial": false,
    "scrapedAt": "2026-08-21T21:21:05.488Z"
}

What ShortMax does not publish — read this before you rely on a field

Three limits are the site's, not this Actor's. They're stated here rather than papered over:

1. follows is derived, not measured. In a full 505-title sweep of the English catalog, 505 of 505 titles had a like count within 1% of exactly one third of their view count (29.5K/88.5K, 23.8K/71.4K, 25.7K/77.1K, 117K/350K…). That is a display formula, not engagement data. Use it as a display value; don't rank on it independently of views. The view counts themselves are the site's own reported figures and may be decorated too.
2. There is no release date. releaseDate is the upload date of the title's cover asset — the closest proxy the site offers — and always ships with releaseDateSource: "cover-asset-path" so it can never be mistaken for a verified date. Episode pages carry a VideoObject.uploadDate, but it is a hardcoded placeholder (2024-03-31T08:00:00+08:00 on every title) and this Actor never uses it.
3. There is no cast. ShortMax publishes no performer credits anywhere; cast is always []. The single Creator: name is surfaced as creator.

Counts are abbreviated on the site, so views is parsed from strings like "88.5K" and is approximate at that magnitude (anything from 88,450 to 88,549). viewsRaw/followsRaw keep the verbatim string so the rounding stays auditable.

Politeness & robots.txt

This Actor scrapes only public catalog metadata. It never touches video sources, HLS manifests, DRM or anything account-related, and it issues zero requests to /search/, which ShortMax disallows in robots.txt. There is deliberately no search mode.

Notes on cost & speed

A verification run over the entire English catalog returned 505 unique titles, zero duplicates and zero failed requests, with views, episodes, genres, category and cover populated on every single record.

ShortMax has no CDN cache in front of its HTML and a multi-second time to first byte (~6 s typical, ~19 s on deep genre pages), with ~200 KB responses. Parallel requests are what make a run fast — the default concurrency of 8 was comfortable in testing with no rate limiting observed. A full 505-title English run is ~527 requests. No proxy is needed by default; the site served plain HTTP clients without cookies or challenges.