Silkline DataZhihu API
Scrape Zhihu (知乎) long-form Chinese Q&A: search answers, articles and pins, all answers of a question, comments and replies, the hot list, user profiles, followers and full answer text. Public author IDs and IP location where shown. No login or cookies needed. Export JSON, CSV or Excel.
Zhihu API in one table
| Official public API for people outside China | No. This tool collects the publicly visible data instead. |
|---|---|
| What you send | A keyword, a link or an ID, depending on the request type |
| What you get back | JSON rows with English field names; 18 request types |
| Login, cookies, Chinese phone number | None needed |
| Output formats | JSON, CSV, Excel, or the Apify API |
| Price | $0.004 per result on the Bronze tier, no subscription |
| Last verified with a real run | 2026-11-09 |
18 kinds of Zhihu request
Zhihu (知乎) is a Chinese question-and-answer website. Users ask questions, write long answers, publish articles and share short posts called pins, and the site keeps a hot list of trending questions. Because answers and articles are often long and detailed, Zhihu is useful for research on Chinese-language opinions and expert text.
- Search by keyword (answers, articles, pins, users, topics, columns, videos, ebooks, papers)
- All answers of questions
- Comments of answers or pins
- Replies to comments
- Hot list
- User profiles
- A user's answers, articles, pins or followed items
- Followers of users
- Accounts users follow
- Full text of answers, articles and pins
- Articles of columns
- Search words: preset and discovery
- Search completions of a keyword
- Paid-column (Salt) content search
- Articles related to an article
- Recommended home feed
- Articles a user is included in
- Accounts Zhihu suggests to follow
A real run, 21 fields per record
Run on 2026-10-26 with input {"mode": "search", "keyword": "人工智能", "searchType": "answers", "sort": "votes", "maxResults": 3}. Personal fields of ordinary users are masked here.
| questionTitle | text | excerpt | voteupCount | commentCount | favoriteCount |
|---|---|---|---|---|---|
| 这波人工智能浪潮怎么越来越像泡沫，AI到底带来了什么？ | 我尽量把事情讲清楚，然后咱们判断一下。 大白话。就那么个意思。不严谨，不科学，不系统，不全面，不保证对，不负任何责... | 我尽量把事情讲清楚，然后咱们判断一下。大白话。就那么个意思。不严谨，不科学，不系统，不全面，不保证对，不负任何责任... | 14127 | 87 | 10058 |
| 如何看待在 2019 年 7 月 3 日的百度人工智能开发者大会上李彦宏遭现场泼水 ？ | 上图： 不对 脖子上挂着牌子穿着有点像工作人员的上台给罗宾头上倒水，表情专注，左手轻轻压住罗宾左手。 罗宾此时... | 上图：不对 | 8823 | 545 | 278 |
All 21 fields in one record
typeansweranswerId2006737249803454088urlhttps://www.zhihu.com/question/658385796/answer/200673724...questionId658385796questionTitle这波人工智能浪潮怎么越来越像泡沫，AI到底带来了什么？text我尽量把事情讲清楚，然后咱们判断一下。 大白话。就那么个意思。不严谨，不科学，不系统，不全面，不保证对，不负任何责...excerpt我尽量把事情讲清楚，然后咱们判断一下。大白话。就那么个意思。不严谨，不科学，不系统，不全面，不保证对，不负任何责任...voteupCount14127commentCount87favoriteCount10058createdAt2026-02-16T06:30:43ZupdatedAt2026-02-22T12:49:04ZauthorauthorIdauthorUrlTokenauthorAvatarauthorHeadlineauthorGenderauthorFollowerCountauthorVerificationauthorIpLocation
Call the Zhihu API
Every run is one HTTP call: send the input as JSON, get the rows back. The same works from Make, Zapier or n8n.
Pay per result
Bronze tier prices from the tool's listing; higher Apify tiers pay less. Free Apify users get two trial runs with their own input.
Full price table and worked examples
The Actor uses pay-per-event pricing:
| Event | Price |
|---|---|
| Run start | $0.00005 per run |
| Result | $0.004 per item returned |
Examples:
- 20 results: $0.00005 + 20 × $0.004 = $0.08005
- 100 results: $0.00005 + 100 × $0.004 = $0.40005
- 1,000 results: $0.00005 + 1,000 × $0.004 = $4.00005
On the free Apify plan you get 2 free trial runs of this Actor with your own input, up to 10 results each, and nothing is charged; after that, runs return a stored sample dataset (a real result of the prefilled input). Use any paid Apify plan for full runs.
The run-start fee is charged once per run, when the first request succeeds, even if that request returns nothing. It is not charged when the run fails from the start or rejects the IDs, or when your spending limit cannot cover it. You never pay for more than maxResults items, and you are never charged for items that were not delivered. Deleted comments and comments without text are skipped and not charged. With includeReplies, each reply counts as one more result. The run stops by itself if your spending limit is reached.
Keyword search through an alternate search path
A keyword search (search types all, answers or articles) may be served by an alternate search path that returns the same records with the same fields, sort order and filters. The other search types (users, topics, columns, videos, ebooks) and the other modes have no alternate path.
Results from the alternate path cost the same $0.004 each, and nothing is added:
- A page that brings no new result, or that times out, is not charged.
- The run-start fee is charged with the first page that delivers a result.
- The search goes on page by page until
maxResultsis reached, there are no more results, or a page brings no new record. A page usually has 13 to 18 results; if pages keep bringing only a few each, the search ends early with what it has, and the run status says so. - Before each page, the run's spending limit must cover the results still wanted from it (at most 20 results, $0.08). If it cannot, the Actor does not request the page and charges nothing for it.
Zhihu API questions
Do I need a Zhihu account, cookies or API keys?
No. You do not need a Zhihu account, API keys, proxies or cookies.
Is it legal to scrape Zhihu?
The Actor returns only public data. Public profile data is included: user ID, profile name, avatar, headline, gender, verification and IP location where Zhihu shows them, for authors, commenters and listed followers alike. Phone numbers, real names and private messages are never part of the output. You are responsible for how you use the data and for following the laws that apply to you.
How fresh is the data?
Data is read from Zhihu when the run happens. Counts such as `voteupCount` or `likeCount` are the values at the time of the run, and the hot list is ranked as on Zhihu at that moment.
Why are some fields empty?
Zhihu does not show every field everywhere. For example, IP location appears only on comments, replies, articles opened with `detail` and profile lookups. See "Fields that are often empty" above.
How many results can I get?
Up to 5,000 per run via `maxResults`. `question` mode returns up to 200 answers per question, and the hot list has about 30 questions per request.
Can I schedule runs and export the data?
Yes. Save your input as a task, add it to a schedule in the Apify Console, and export each run's dataset as JSON, CSV, Excel or other formats. The steps are in [Large batches and scheduled runs](#large-batches-and-scheduled-runs).