Douban Scraper - Movies, TV, Books, Music & Reviews

Scrape Douban - China's largest film, TV, book and music rating community - into clean CSV/JSON. Ratings with the full 5-star distribution, complete cast and crew, awards, IMDb IDs, short comments, long reviews and people profiles.

No login. No cookies to supply. No browser.

---

Why this actor exists (demand receipt)

Measured against the live Apify Store API on 2026-07-22, not a stale report:

Signal	Value
Douban actors on the Apify Store	10
Combined lifetime users	116
Combined 30-day users	61
Largest incumbent	zhorex/douban-scraper - 57 users, 39 monthly

61 of 116 lifetime users were active in the last 30 days - a 53% active ratio, the healthiest of any China/APAC target still unbuilt. For comparison on the same day: Temu 285/2,680 = 10.6%, Shein 145/949 = 15%, Lazada 137/1,088 = 12.6%. Douban demand is current, not a decayed lifetime total.

Every incumbent covers a single slice - movies only, or reviews only. None covers movies + TV + books + music + comments + reviews + people in one actor, and none returns the rating distribution, the full cast, awards or the IMDb ID.

---

What you get

Mode	One record per	Highlights
Movies & TV	film / series	rating + full star distribution, genres, countries, languages, every regional release date, runtime, IMDb ID, synopsis, alternate titles, directors/writers/cast with person IDs, full cast & crew (60+ people), awards with win/nomination status, similar-title recommendations, want-to-watch & watched counts
Books	book	authors, translators, publisher, imprint, publication date, pages, binding, ISBN, series, price, rating + distribution, content summary, author biography, table of contents
Music albums	album	performers, release date, genre, album type, medium, label, barcode, ISRC, full track listing, rating + distribution
Short comments	comment	text, star rating, author + profile URL + avatar, timestamp, useful votes, IP region, permalink
Long reviews	review	headline, full body, star rating, author, timestamp, useful / not-useful votes, reply count, permalink
People	person	Chinese + original name, gender, date of birth, birthplace, occupations, alternate names, family members, IMDb person ID, biography, follower count, best-known titles

One mode returns exactly one record type, so your CSV never columnises on a union of unrelated keys.

Sample (Movies & TV, trimmed)

{
  "type": "movie",
  "subjectId": "1292052",
  "url": "https://movie.douban.com/subject/1292052/",
  "titleChinese": "肖申克的救赎",
  "titleOriginal": "The Shawshank Redemption",
  "akaTitles": ["月黑高飞(港)", "刺激1995(台)", "地狱诺言", "铁窗岁月"],
  "year": 1994,
  "isTvSeries": false,
  "rating": 9.7,
  "ratingVotes": 3306439,
  "ratingDistribution": { "star5": 88.3, "star4": 11.0, "star3": 0.6, "star2": 0.1, "star1": 0.1 },
  "ratingBetterThan": ["99% 的犯罪片", "99% 的剧情片"],
  "genres": ["剧情", "犯罪"],
  "countries": ["美国"],
  "languages": ["英语"],
  "releaseDates": ["1994-09-10(多伦多电影节)", "1994-10-14(美国)"],
  "runtimeMinutes": 142,
  "imdbId": "tt0111161",
  "imdbUrl": "https://www.imdb.com/title/tt0111161/",
  "directors": [{ "name": "弗兰克·德拉邦特", "personId": "1047973" }],
  "cast": [{ "name": "蒂姆·罗宾斯 Tim Robbins", "personId": "27260288", "role": "演员 Actor", "knownFor": ["肖申克的救赎", "神秘河"] }],
  "awards": [{ "award": "第67届奥斯卡金像奖", "year": 1995,
               "categories": [{ "category": "最佳影片(提名)", "won": false, "people": ["妮基·马文"] }] }],
  "rank": 1
}

---

Quick start

Douban Movie Top 250, fully enriched - no URLs needed:

{ "mode": "movies", "collection": "top250", "maxItems": 250 }

Search for science-fiction titles:

{ "mode": "movies", "searchQueries": ["科幻"], "maxItems": 100 }

Everything about one title:

{ "mode": "movies", "startUrls": [{ "url": "https://movie.douban.com/subject/1292052/" }] }

Its 500 most useful comments:

{ "mode": "comments",
  "startUrls": [{ "url": "https://movie.douban.com/subject/1292052/" }],
  "maxItems": 500 }

Its entire cast as person profiles - pass a movie URL in People mode:

{ "mode": "people", "startUrls": [{ "url": "https://movie.douban.com/subject/1292052/" }], "maxItems": 60 }

Douban Book Top 250 with ISBNs:

{ "mode": "books", "collection": "top250", "maxItems": 250 }

---

Input reference

Field	Type	Default	Notes
mode	enum	movies	movies · books · music · comments · reviews · people
collection	enum	none	top250 (movies/books/music) · chart · nowPlaying · tag
searchQueries	string[]	[]	Chinese or English keywords. Movies & Books only - see note below
startUrls	url[]	[]	Subject or person URLs
tag	string	-	Used with collection: "tag"
city	string	beijing	Used with collection: "nowPlaying"
maxItems	integer	100	0 = no cap (the wall-clock budget stops the run)
includeDetails	boolean	true	Opens each subject page for the rich fields
includeCast	boolean	true	Full cast & crew (movies)
includeAwards	boolean	true	Awards with win/nomination status (movies)
maxPerSubject	integer	100	Comments/reviews per subject
minRating	integer	-	Keep only records rated ≥ this (0-10)
detailConcurrency	integer	4	Parallel detail fetches, each on its own IP
proxyConfiguration	proxy	Residential	Strongly recommended - Douban throttles per IP

---

Field coverage - measured, not claimed

From a real platform run (Movie Top 250, details + cast + awards): 38 of 41 movie fields populated on ≥70% of rows.

Fields that are legitimately sparse, and why:

Field	Why
episodeCount, episodeRuntimeMinutes	Television only. Douban does not print an episode count on a feature film - these populate on series and stay null on films, which is the correct answer, not a gap.
sourceQuery	Provenance. Set only when a record came from a search term; null for chart/collection/URL runs.
rank	Only Top 250 pages number their entries. Chart, now-playing, tag and search results have no rank.
tagline	The one-line quote is a Top 250 feature; subject pages do not carry it.
location (comments)	Douban only shows a poster's IP region on recent comments; historic ones predate the feature.
watchStatus (book/music comments)	Only film and TV comments carry a 看过/在看 badge; book and album comments do not render one.
awards, cast	Present when includeAwards / includeCast are on (default) and the title actually has them - documentaries and shorts often have no awards page.

Music has no keyword search. Douban's music suggest endpoint answers 200 with an empty array for every query tried, and the old HTML search page is retired site-wide (it returns an 11,502-byte shell with an empty result list on 10/10 fresh residential sessions across five countries, 2026-07-22). Movie and book search therefore run against Douban's live JSON discovery endpoints instead; in Music mode use Top 250, a tag, or album URLs, and the run says so in its status message rather than silently returning nothing.

There is also deliberately no comment-sort option. Douban's newest-first ordering (sort=time) is login-gated - measured on 5 fresh residential sessions on 2026-07-22 it returned 403 five times out of five, while the most-useful ordering returned 20 comments on 3 of the same 5. A switch that always fails is worse than no switch, so comments always come back in Douban's usefulness order.

Two fields are deliberately not shipped rather than shipped empty: Douban stops rendering the 常用标签 tag cloud and a person's filmography to anonymous callers (verified against full 117 KB / 101 KB / 25 KB page dumps on 2026-07-22), so no always-blank column is created. Use a movie record's cast array, or People mode with a movie URL, to get filmography-style links.

---

How it works

Douban's subject pages do not serve a CAPTCHA. They redirect to sec.douban.com which returns a small SHA-512 proof-of-work challenge: find a nonce whose sha512(challenge + nonce) begins with four zeros. That is roughly 90 ms of CPU in Node - no browser, no solver service, no cost.

The actor solves it, posts the answer back, and the resulting cookie unlocks every subsequent request on that session. Because Douban rate-limits per IP, sessions are pooled: each one carries its own cookie jar on its own residential exit, requests round-robin across the pool, and only the individual session that gets throttled is replaced.

Practical consequences for you:

• No browser - so runs are fast and cheap, and no browser binary can break the build.
• Transient blocks never fail a run. A throttled run ends Succeeded with a status message telling you to re-run, and pay-per-event means you are charged $0 for records you did not receive.
• It cannot time out. The actor reads its own run deadline, stops starting new work before it, emits everything already collected, and exits Succeeded - so no input size or slow response can produce a TIMED-OUT run.

---

Use with AI assistants (MCP)

This actor is available through the Apify MCP server, so assistants like Claude can call it directly:

"Get the Douban Top 250 movies with their ratings and IMDb IDs."

Agent-ready: autonomous payments (x402 & Skyfire)

This actor is agent-ready — AI agents can discover it, run it, and pay for it autonomously, with no Apify account and no human in the loop. It uses pay-per-event pricing and limited permissions, so it qualifies for Apify's agentic-payment standards:

• x402 — an open, HTTP-native payment protocol. Agents pay per run in USDC on the Base network directly through the Apify MCP server — no account, no API key.
• Skyfire — agent-to-service payments for fully autonomous AI-agent workflows.

Building an AI agent, MCP tool, or autonomous data pipeline? This scraper is ready to plug in and pay as it goes.

---

Pricing

Pay-per-event - you pay for records delivered, never for start-up, pagination or a blocked attempt.

Event	Price
Movie/TV full record	$0.006
Movie/TV listing (details off)	$0.003
Book full record	$0.006
Book listing (details off)	$0.003
Music album full record	$0.005
Music album listing (details off)	$0.003
Person profile	$0.005
Long review	$0.002
Short comment	$0.001

---

Legal

Scrapes publicly available pages only - no login, no paywall, no personal data behind authentication. You are responsible for how you use the output, including Douban's terms and any applicable copyright or data-protection law. Review text and user names are user-generated content owned by their authors.