Zhihu Scraper — Posts, Pages, Comments, Hot List

Zhihu Scraper (知乎) — Search, Answers & Articles, Comments, Hot List

Scrape Zhihu (知乎), China's largest Q&A and long-form knowledge platform, in one actor:

• 🔎 Keyword search — answers, articles and videos for any Chinese or English keyword, with content-type, sort and date filters (last 24 h, week, month…)
• 📄 Full pages — the complete text of an answer, article or video page, plus its comments and comment replies
• 👤 Author articles — every article from a profile link, plus the profile's answer and article counts
• 🔥 Hot list (热榜) — the 30 trending questions with heat score (e.g. 1095 万热度 → 10,950,000), answer, comment and follower counts

No login, no cookies, no browser setup. Export to JSON, CSV, Excel or straight into your pipeline via the Apify API.

Who uses it

• China market research — what Chinese consumers and professionals say about a brand, product or industry (新能源汽车, 光伏, AI…)
• Trend monitoring — daily snapshot of the Zhihu hot list next to Weibo, Douyin and Baidu
• Social listening & sentiment — comments and replies under the pages that matter to you
• Content and KOL research — an author's full article history with likes, comments and favourites
• AI / LLM datasets — long-form Chinese text with clean metadata and provenance

Modes

Mode	What you get	Rows	Price per row
search	answers, articles, videos by keyword; type, sort and date filters	20–100 per keyword	$0.0028
hot	热榜 hot list with heat score, answer, comment and follower counts	up to 30	$0.0008
content	full page (complete text) and, optionally, comments and replies	1 + up to 100 comments	$0.0028 page, $0.0004 comment
user	profile counters and the author's articles	1 + up to 100	$0.001 profile, $0.0028 article
all	everything above in one run	—	—

Plus $0.009 per run start. Prices shown are for the FREE plan; paid Apify plans get 10–30% off (BRONZE −10%, SILVER −20%, GOLD −30%). 1,000 search results ≈ $2.80. Empty answers from Zhihu are never charged.

Input examples

Search by keyword, articles only, most upvoted first:

{
  "mode": "search",
  "keywords": ["新能源汽车", "光伏"],
  "contentType": "article",
  "sort": "upvote_count_descending",
  "postedWithin": "year",
  "limitPerQuery": 50
}

A page with its comments and replies:

{
  "mode": "content",
  "contentUrls": ["https://zhuanlan.zhihu.com/p/714800391"],
  "includeComments": true,
  "includeCommentReplies": true
}

An author's articles:

{ "mode": "user", "userUrls": ["https://www.zhihu.com/people/jaliou"] }

Today's hot list:

{ "mode": "hot" }

Output example

{
  "type": "zhihu-content",
  "id": "714800391",
  "contentType": "article",
  "title": "硬核干货！2026年最新新能源汽车行业411篇顶级报告，吐血整理！…",
  "excerpt": "包括小米汽车、理想汽车、比亚迪、蔚来、零跑汽车、长城汽车、特斯拉、华为汽车…",
  "fullText": null,
  "url": "https://zhuanlan.zhihu.com/p/714800391",
  "likeCount": 231,
  "commentCount": 11,
  "collectCount": 0,
  "postedAt": "2024-08-16T02:03:48+00:00",
  "coverImageUrl": null,
  "author": {
    "id": "61a5c0552f567bdaca1a4125d85b4c76",
    "name": "小狮子",
    "url": "https://www.zhihu.com/people/jaliou",
    "followerCount": 4979,
    "verified": false
  },
  "keyword": "新能源汽车",
  "foundVia": "search",
  "scrapedAt": "2026-11-08T12:52:11.659Z"
}

Every row has a type field:

• zhihu-content — id, contentType, title, excerpt, fullText, url, likeCount, commentCount, collectCount, postedAt, coverImageUrl, author{…}, keyword, foundVia, provenance
• zhihu-hot — rank, id, title, url, excerpt, questionType, answerCount, commentCount, followerCount, heat, heatText, createdAt, thumbnailUrl
• zhihu-comment — id, text, likeCount, replyCount, parentCommentId, author{…}, contentUrl
• zhihu-profile — url, username, answerCount, articleCount, availableFields
• notice / error — free service rows explaining what was skipped and why; filter them out with type != "notice"

fullText is filled in content mode; search and author rows carry the excerpt. Each row also has a provenance block (source, canonical URL, collected-at time, checksum) for audit trails.

Good to know

• Zhihu search is flaky by nature: the first answer is often empty. The actor retries on its own (retries, default 8, about 2 s each) — empty answers cost nothing. If search stays empty, an optional site:zhihu.com web search fallback returns title, link and snippet rows marked foundVia: web-fallback.
• Hot list heat is Zhihu's own 热度 figure converted to a number: heatText keeps the original (1095 万热度), heat holds 10950000.
• Posted within filters by publication date on our side (Zhihu search has no date filter). Dropped rows are not charged; combine with Newest first for the freshest results.
• Question pages are not supported in content mode: Zhihu closes the answer list of a question. Use links to individual answers (/question/<id>/answer/<id>), articles or videos. Unsupported links are skipped with a free notice row.
• Profiles return only the answer and article counts (availableFields says so); the author's articles come in full.
• Max cost per run is respected: rows beyond your limit are not written and not charged.
• A run never fails because of the input: missing or wrong input gives a free notice row, and if nothing at all can be collected, you get the hot list.
• At the end of each run a free summary row and a SUMMARY record in the key-value store show what was delivered and charged.

FAQ

Do I need a Zhihu account or cookies? No.

Can I search in English? Yes, but Zhihu content is Chinese — Chinese keywords give far more results.

How fresh is the data? Everything is fetched live at run time. Schedule the actor (e.g. hourly hot list) to build a time series.

Is scraping Zhihu legal? The actor collects only publicly visible data. Make sure your use complies with applicable laws and Zhihu's terms, especially for personal data.

Related

• Need only trending lists across platforms? China Hot Lists Scraper (apimy/china-hot-lists-scraper) puts the Baidu, Zhihu, Weibo, Douyin and Bilibili hot lists into one dataset.
• More Chinese platforms from the same author: Bilibili Scraper, Douyin, RedNote (Xiaohongshu) and Weibo scrapers.