Pricing

from $1.80 / 1,000 content matches

Try for free

Go to Apify Store

Zhihu Scraper (知乎)

Try for free

Search public Zhihu by keyword, or open questions, articles, columns, videos, comments, and creator pages. Get structured records with source links, available text, authors, engagement, dates, and related public context.

Pricing

from $1.80 / 1,000 content matches

Rating

0.0

(0)

Developer

Maxime Dupré

Actor stats

Bookmarked

Total users

Monthly active users

a day ago

Last modified

🔎 Find public Zhihu research records

For researchers, product teams, marketers, and developers, this Zhihu scraper searches public Zhihu or opens known public pages. It returns structured records with source links, available text, authors, engagement counts, dates, and links between questions, answers, articles, comments, columns, media, and profiles. Use the rows to review public discussions, compare topics, find creators, or export research data.

Search public Zhihu topics about stocks with Zhihu stock.
Review public Zhihu discussions about Douyin.
Track the current Zhihu hot list with rank, heat, and top-answer details.
Collect top-level comments from a public Zhihu question, answer, article, or video.
Search public Zhihu discussions about Chinese to English.

📚 Explore Zhihu research records

The dataset has nine record types: answers, questions, hot questions, articles, column articles, videos, comments, comment replies, and creator profiles. Each row has a recordType and a required source.url. Other fields appear when the public source provides them.

The Actor saves the first eligible occurrence of a source item as soon as it finds it. If the same item appears later through another keyword, URL, handle, or ID, the later match is ignored and does not rewrite the saved row. source.discoveryInput shows the one input that first led to the saved row.

Missing public values are left out. When the source does not expose a data area, sourceGaps can name that gap. Public source text, identifiers, and URLs stay as returned by Zhihu.

▶️ Choose a Zhihu collection

Choose one result type (target) in the Input tab. hotQuestions needs no other input.
For keywordSearch, enter one keyword or phrase and choose the content types, filters, and order.
For another result type, add the matching public URL or URLs, such as question, article, column, video, content, comment, or creator URLs.
Set maxItems when you want fewer rows to review. The form starts at 10. Leave it empty or set 0 to get all available results.
Run the Actor and open the default dataset in the run's Storage tab.

Use another run for a different keyword. The Actor handles public pages only. It does not cover private, deleted, login-only, paywalled, or otherwise restricted Zhihu content. It does not write to Zhihu or create summaries, translations, sentiment labels, or other derived analysis.

⚙️ Input

Public input example

This is the public input from a successful default-input run:

{
  "maxItems": 10,
  "target": "keywordSearch",
  "keyword": "machine learning",
  "contentTypes": [
    "answers"
  ],
  "sortBy": "relevance"
}

Input fields

Field	Type	What it does
`target`	string, required	Chooses one result type: `keywordSearch`, `hotQuestions`, `questionAnswers`, `article`, `columnArticles`, `video`, `comments`, `commentReplies`, or `creatorProfile`. Values in other result type sections are ignored.
`keyword`	string	Searches public Zhihu with one keyword or phrase. Use another run for a different search.
`contentTypes`	array of strings	Limits keyword search to `answers`, `questions`, `articles`, `videos`, `columns`, or `people`. Leave it empty to search all supported types.
`topicCategories`	array of strings	Keeps keyword results with the listed topic categories. Leave it empty to use all categories.
`hotOnly`	boolean	Keeps only keyword results that belong to the current Zhihu hot list.
`minimumTopAnswerVotes`	integer	Keeps keyword results whose top answer has at least this many votes. Leave it empty to skip this filter.
`publishedFrom`	string, `YYYY-MM-DD`	Keeps keyword results published on or after this UTC date. Leave it empty for no start date.
`publishedTo`	string, `YYYY-MM-DD`	Keeps keyword results published on or before this UTC date. Leave it empty for no end date.
`sortBy`	string	Orders keyword results by `relevance`, `newest`, or `votes`.
`questionUrls`	array of objects	Supplies one or more public Zhihu question URLs for answer collection.
`questionUrls[].url`	string	A public Zhihu question URL.
`articleUrls`	array of objects	Supplies one or more public Zhihu article URLs for article detail collection.
`articleUrls[].url`	string	A public Zhihu article URL.
`columnUrls`	array of objects	Supplies one or more public Zhihu column URLs for column article listings.
`columnUrls[].url`	string	A public Zhihu column URL.
`videoUrls`	array of objects	Supplies one or more public Zhihu video URLs for video collection.
`videoUrls[].url`	string	A public Zhihu video URL.
`contentUrls`	array of objects	Supplies one or more public Zhihu question, answer, article, or video URLs for top-level comment collection.
`contentUrls[].url`	string	A public Zhihu question, answer, article, or video URL.
`commentSort`	string	Orders top-level comments by `popular` or `newest`. Leave it empty for Zhihu's normal comment order.
`commentUrls`	array of objects	Supplies one or more public Zhihu comment URLs for reply collection.
`commentUrls[].url`	string	A public Zhihu comment URL.
`creatorUrls`	array of objects	Supplies one or more public Zhihu creator profile URLs. The Actor returns the profile details for each creator.
`creatorUrls[].url`	string	A public Zhihu creator profile URL.
`maxItems`	integer	Stops after this many records. The form starts at 10. Leave it empty or set 0 to get all available results.

🧾 Output

Output fields

The tables below list the complete public fields for each record shape. Dotted paths describe fields inside an object or an array item. A field is omitted when Zhihu does not expose it for that row. Types use plain names such as string, integer, number, boolean, object, and array.

Answer row

Field	Type	What it does
`recordType`	string	Identifies the row as an `answer`.
`source`	object	Holds the public answer page and its source identity.
`source.id`	string	Gives the answer ID when Zhihu exposes one.
`source.url`	string URL	Gives the public answer URL.
`source.discoveryInput`	string	Gives the keyword, URL, handle, or ID that first found the answer.
`title`	string	Gives a public title when available.
`summary`	string	Gives the public excerpt when available.
`text`	string	Gives the available public answer body.
`author`	object	Holds public answer-author details.
`author.id`	string	Gives the author ID when available.
`author.name`	string	Gives the public author name.
`author.url`	string URL	Gives the public author profile URL.
`author.headline`	string	Gives the public author headline when available.
`author.isVerified`	boolean	Says whether Zhihu marks the author as verified.
`engagement`	object	Holds public engagement counts.
`engagement.voteCount`	integer	Gives the public vote count when available.
`engagement.likeCount`	integer	Gives the public like count when available.
`engagement.commentCount`	integer	Gives the public comment count when available.
`engagement.replyCount`	integer	Gives the public reply count when available.
`engagement.viewCount`	integer	Gives the public view count when available.
`engagement.favoriteCount`	integer	Gives the public favorite count when available.
`engagement.shareCount`	integer	Gives the public share count when available.
`publishedAt`	date-time string	Gives the UTC publication time when available.
`createdAt`	date-time string	Gives the UTC creation time when available.
`updatedAt`	date-time string	Gives the UTC update time when available.
`question`	object	Holds the public question linked to the answer.
`question.id`	string	Gives the linked question ID when available.
`question.title`	string	Gives the linked question title.
`question.url`	string URL	Gives the linked question URL.
`topics`	array of objects	Lists public topics linked to the answer.
`topics[].id`	string	Gives a topic ID when available.
`topics[].name`	string	Gives a topic name.
`topics[].url`	string URL	Gives a topic URL when available.
`sourceGaps`	array of strings	Lists public data areas not exposed for the row.

Genuine answer row

{
  "recordType": "answer",
  "source": {
    "url": "https://www.zhihu.com/question/279973545/answer/444230992",
    "id": "444230992",
    "discoveryInput": "machine learning"
  },
  "text": "<p data-pid=\"9tGfa17N\">人工智能并不是一个新的术语，它已经有几十年的历史了，大约从80年代初开始，计算机科学家们开始设计可以学习和模仿人类行为的算法。</p><p data-pid=\"IJh2oos6\">在算法方面，最重要的算法是神经网络，由于过拟合而不是很成功（模型太强大，但数据不足）。尽管如此，在一些更具体的任务中，使用数据来适应功能的想法已经取得了显着的成功，并且这也构成了当今机器学习的基础。</p><p data-pid=\"qrDY5iV4\">在模仿方面，人工智能专注于图像识别，语音识别和自然语言处理。人工智能专家们花费了大量的时间来创建诸如边缘检测，颜色配置文件，N-gram，语法树等。不过，这些进步还不足以达到我们的需求。</p><p data-pid=\"QhTWVTSH\"><b>传统的机器学习：</b></p><p data-pid=\"wsf2CpJ3\">机器学习（ML）技术在预测中发挥了重要的作用，ML经历了多代的发展，形成了具有丰富的模型结构，例如：</p><p data-pid=\"eei0zSnp\">1.线性回归。</p><p data-pid=\"FnKTFjAI\">2.逻辑回归。</p><p data-pid=\"EOVpRjoj\">3.决策树。</p><p data-pid=\"VOagVVvL\">4.支持向量机。</p><p data-pid=\"b7unDNo2\">5.贝叶斯模型。</p><p data-pid=\"dK3BSba2\">6.正则化模型。</p><p data-pid=\"FwFwZ8tB\">7.模型集成（ensemble）。</p><p data-pid=\"B3Ng-01L\">8.神经网络。</p><p data-pid=\"EJHNNAie\">这些预测模型中的每一个都基于特定的算法结构，参数都是可调的。训练预测模型涉及以下步骤：</p><p data-pid=\"9_cfmntE\">1. <b>选择一个模型结构</b>（例如逻辑回归，随机森</p>",
  "author": {
    "name": "阿里云云栖号",
    "id": "65b52731faea3de2e983890f3f3fffa1",
    "url": "https://www.zhihu.com/people/a-li-yun-yun-qi-she-qu-48",
    "headline": "汇聚2026云栖大会最新优质内容，了解最新科技动态！",
    "isVerified": true
  },
  "engagement": {
    "voteCount": 251,
    "commentCount": 4
  },
  "createdAt": "2018-07-16T14:52:19.000Z",
  "updatedAt": "2018-07-16T14:52:19.000Z",
  "question": {
    "id": "279973545",
    "title": "机器学习、深度学习和强化学习的关系和区别是什么？",
    "url": "https://www.zhihu.com/question/279973545"
  },
  "topics": [
    {
      "name": "人工智能",
      "id": "19551275",
      "url": "https://www.zhihu.com/topic/19551275"
    },
    {
      "name": "算法",
      "id": "19553510",
      "url": "https://www.zhihu.com/topic/19553510"
    },
    {
      "name": "机器学习",
      "id": "19559450",
      "url": "https://www.zhihu.com/topic/19559450"
    },
    {
      "name": "强化学习 (Reinforcement Learning)",
      "id": "20039099",
      "url": "https://www.zhihu.com/topic/20039099"
    },
    {
      "name": "AI技术",
      "id": "20106982",
      "url": "https://www.zhihu.com/topic/20106982"
    }
  ],
  "sourceGaps": [
    "fullText"
  ],
  "summary": "人工智能并不是一个新的术语，它已经有几十年的历史了，大约从80年代初开始，计算机科学家们开始设计可以学习和模仿人类行为的算法。 在算法方面，最重要的算法是神经网络，由于过拟合而不是很成功（模型太强大，但数据不足）。尽管如此，在一些更具体的任务中，使用数据来适应功能的想法已经取得了显着的成功，并且这也构成了当今机器学习的基础。 在模仿方面，人工智能专注于图像识别，语音识别和自然语言处理。人工智能专家们…"
}

Question row

Field	Type	What it does
`recordType`	string	Identifies the row as a `question`.
`source`	object	Holds the public question page and its source identity.
`source.id`	string	Gives the question ID when available.
`source.url`	string URL	Gives the public question URL.
`source.discoveryInput`	string	Gives the keyword, URL, handle, or ID that first found the question.
`title`	string	Gives the public question title.
`summary`	string	Gives a public question excerpt when available.
`text`	string	Gives public question text when available.
`author`	object	Holds public question-author details.
`author.id`	string	Gives the author ID when available.
`author.name`	string	Gives the public author name.
`author.url`	string URL	Gives the public author profile URL.
`author.headline`	string	Gives the public author headline when available.
`author.isVerified`	boolean	Says whether Zhihu marks the author as verified.
`engagement`	object	Holds public engagement counts.
`engagement.voteCount`	integer	Gives the public vote count when available.
`engagement.likeCount`	integer	Gives the public like count when available.
`engagement.commentCount`	integer	Gives the public comment count when available.
`engagement.replyCount`	integer	Gives the public reply count when available.
`engagement.viewCount`	integer	Gives the public view count when available.
`engagement.favoriteCount`	integer	Gives the public favorite count when available.
`engagement.shareCount`	integer	Gives the public share count when available.
`publishedAt`	date-time string	Gives the UTC publication time when available.
`createdAt`	date-time string	Gives the UTC creation time when available.
`updatedAt`	date-time string	Gives the UTC update time when available.
`topics`	array of objects	Lists public topics linked to the question.
`topics[].id`	string	Gives a topic ID when available.
`topics[].name`	string	Gives a topic name.
`topics[].url`	string URL	Gives a topic URL when available.
`sourceGaps`	array of strings	Lists public data areas not exposed for the row.

Genuine question row

{
  "recordType": "question",
  "source": {
    "url": "https://www.zhihu.com/question/1927140506573435010",
    "id": "1927140506573435010",
    "discoveryInput": "machine learning"
  },
  "title": "Kimi 发布首个万亿参数开源模型 K2 模型，哪些信息值得关注？",
  "author": {
    "name": "holiday111",
    "id": "a1cdaa3353c17f8cc92042efd18d3da1",
    "url": "https://www.zhihu.com/people/hongdi-10",
    "headline": "菜鸡但爱学，窝囊但会发疯",
    "isVerified": false
  },
  "engagement": {
    "voteCount": 91,
    "commentCount": 5
  },
  "createdAt": "2025-07-11T15:01:41.000Z",
  "updatedAt": "2025-07-11T15:51:13.000Z",
  "topics": [
    {
      "name": "人工智能",
      "id": "19551275",
      "url": "https://www.zhihu.com/topic/19551275"
    },
    {
      "name": "SOTA",
      "id": "25520717",
      "url": "https://www.zhihu.com/topic/25520717"
    },
    {
      "name": "开源模型",
      "id": "27782131",
      "url": "https://www.zhihu.com/topic/27782131"
    },
    {
      "name": "Kimi智能助手",
      "id": "29603811",
      "url": "https://www.zhihu.com/topic/29603811"
    },
    {
      "name": "MoE优化",
      "id": "926442969",
      "url": "https://www.zhihu.com/topic/926442969"
    }
  ],
  "summary": "7 月 11 日， Kimi K2 模型发布并同步开源，模型及 fp8 权重文件已开源至 Hugging Face https://huggingface.co/moonshotai/Kimi-K2-Instruct [文章: Kimi K2 发布并开源，擅长代码与 Agentic 任务]"
}

Hot question row

Field	Type	What it does
`recordType`	string	Identifies the row as a `hotQuestion`.
`source`	object	Holds the public hot-question page and its source identity.
`source.id`	string	Gives the question ID when available.
`source.url`	string URL	Gives the public question URL.
`source.discoveryInput`	string	Shows the input that first found the hot question.
`title`	string	Gives the hot-question title.
`summary`	string	Gives the public hot-question excerpt when available.
`text`	string	Gives public question text when available.
`author`	object	Holds public question-author details.
`author.id`	string	Gives the author ID when available.
`author.name`	string	Gives the public author name.
`author.url`	string URL	Gives the public author profile URL.
`author.headline`	string	Gives the public author headline when available.
`author.isVerified`	boolean	Says whether Zhihu marks the author as verified.
`engagement`	object	Holds public engagement counts.
`engagement.voteCount`	integer	Gives the public vote count when available.
`engagement.likeCount`	integer	Gives the public like count when available.
`engagement.commentCount`	integer	Gives the public comment count when available.
`engagement.replyCount`	integer	Gives the public reply count when available.
`engagement.viewCount`	integer	Gives the public view count when available.
`engagement.favoriteCount`	integer	Gives the public favorite count when available.
`engagement.shareCount`	integer	Gives the public share count when available.
`publishedAt`	date-time string	Gives the UTC publication time when available.
`createdAt`	date-time string	Gives the UTC creation time when available.
`updatedAt`	date-time string	Gives the UTC update time when available.
`topics`	array of objects	Lists public topics linked to the question.
`topics[].id`	string	Gives a topic ID when available.
`topics[].name`	string	Gives a topic name.
`topics[].url`	string URL	Gives a topic URL when available.
`hot`	object	Holds current hot-list context.
`hot.rank`	integer	Gives the hot-list position.
`hot.heatScore`	number	Gives the public heat score when available.
`hot.topAnswer`	object	Holds top-answer context when the hot-list source exposes it.
`hot.topAnswer.id`	string	Gives the top-answer ID when available.
`hot.topAnswer.url`	string URL	Gives the top-answer URL when available.
`hot.topAnswer.textExcerpt`	string	Gives the top-answer excerpt when available.
`hot.topAnswer.author`	object	Holds top-answer author details.
`hot.topAnswer.author.id`	string	Gives the top-answer author ID when available.
`hot.topAnswer.author.name`	string	Gives the top-answer author name.
`hot.topAnswer.author.url`	string URL	Gives the top-answer author URL when available.
`hot.topAnswer.voteCount`	integer	Gives the top-answer vote count when available.
`sourceGaps`	array of strings	Lists public data areas not exposed for the row.

Genuine hot question row

{
  "recordType": "hotQuestion",
  "source": {
    "url": "https://www.zhihu.com/question/2100890499762446584",
    "id": "2100890499762446584",
    "discoveryInput": "hotQuestions"
  },
  "title": "武汉一小区爆燃致 2 死 1 伤，调查、处置工作正在进行中，目前情况如何？",
  "author": {
    "name": "知乎用户",
    "id": "d371683fd9673e521609ae7a4788e856",
    "url": "https://www.zhihu.com/api/v4/people/0",
    "isVerified": false
  },
  "engagement": {
    "voteCount": 0,
    "commentCount": 1
  },
  "createdAt": "2026-11-03T02:02:07.000Z",
  "updatedAt": "2026-11-03T02:02:59.000Z",
  "topics": [
    {
      "name": "社会",
      "id": "19566933",
      "url": "https://www.zhihu.com/topic/19566933"
    },
    {
      "name": "武汉",
      "id": "19570564",
      "url": "https://www.zhihu.com/topic/19570564"
    },
    {
      "name": "爆炸",
      "id": "19599602",
      "url": "https://www.zhihu.com/topic/19599602"
    },
    {
      "name": "燃气爆炸",
      "id": "28113516",
      "url": "https://www.zhihu.com/topic/28113516"
    },
    {
      "name": "爆燃事故",
      "id": "29559166",
      "url": "https://www.zhihu.com/topic/29559166"
    }
  ],
  "summary": "11月3日，武汉市公安局武昌区分局发布警情通报： 11月2日23时50分许，武昌区一小区发生爆燃，造成2人死亡，1人受伤(无生命危险)。公安、消防救援、120急救等部门迅速赶到现场处置，开展伤员救治、现场勘查等工作。目前，调查、处置工作正在进行中。 [图片] 武汉市武昌区一小区爆燃致2死1伤_腾讯新闻",
  "hot": {
    "rank": 2,
    "heatScore": 530,
    "topAnswer": {
      "id": "2101050722754340774",
      "url": "https://www.zhihu.com/question/2100890499762446584/answer/2101050722754340774",
      "textExcerpt": "明明是bfx社会爆炸新闻，结果一点热度没有 [图片]",
      "author": {
        "name": "闲在家的狗子",
        "id": "5748162b7fd1016ae0ae667c94a9a851",
        "url": "https://www.zhihu.com/people/zhang-han-wen-12"
      },
      "voteCount": 167
    }
  }
}

Article row

Article detail runs use this record shape.

Field	Type	What it does
`recordType`	string	Identifies the row as an `article`.
`source`	object	Holds the public article page and its source identity.
`source.id`	string	Gives the article ID when available.
`source.url`	string URL	Gives the public article URL.
`source.discoveryInput`	string	Shows the URL, keyword, handle, or ID that first found the article.
`title`	string	Gives the public article title.
`summary`	string	Gives the public article excerpt when available.
`text`	string	Gives the available public article body.
`author`	object	Holds public article-author details.
`author.id`	string	Gives the author ID when available.
`author.name`	string	Gives the public author name.
`author.url`	string URL	Gives the public author profile URL.
`author.headline`	string	Gives the public author headline when available.
`author.isVerified`	boolean	Says whether Zhihu marks the author as verified.
`engagement`	object	Holds public engagement counts.
`engagement.voteCount`	integer	Gives the public vote count when available.
`engagement.likeCount`	integer	Gives the public like count when available.
`engagement.commentCount`	integer	Gives the public comment count when available.
`engagement.replyCount`	integer	Gives the public reply count when available.
`engagement.viewCount`	integer	Gives the public view count when available.
`engagement.favoriteCount`	integer	Gives the public favorite count when available.
`engagement.shareCount`	integer	Gives the public share count when available.
`publishedAt`	date-time string	Gives the UTC publication time.
`createdAt`	date-time string	Gives the UTC creation time when available.
`updatedAt`	date-time string	Gives the UTC update time when available.
`column`	object	Holds the public column linked to the article.
`column.id`	string	Gives the column ID when available.
`column.name`	string	Gives the column name when available.
`column.url`	string URL	Gives the column URL when available.
`topics`	array of objects	Lists public topics linked to the article.
`topics[].id`	string	Gives a topic ID when available.
`topics[].name`	string	Gives a topic name.
`topics[].url`	string URL	Gives a topic URL when available.
`sourceGaps`	array of strings	Lists public data areas not exposed for the row.

Genuine article row

{
  "recordType": "article",
  "source": {
    "url": "https://zhuanlan.zhihu.com/p/682332617",
    "id": "682332617",
    "discoveryInput": "https://zhuanlan.zhihu.com/p/682332617"
  },
  "title": "【基础】大模型的背景知识：AI 宏观了解",
  "author": {
    "name": "张强",
    "id": "387cefbf2c35f338d84d71348436fee2",
    "url": "https://www.zhihu.com/people/studyisalifestyle",
    "headline": "坚持一些事，是因为不想让世界对我全面掌控……",
    "isVerified": false
  },
  "engagement": {
    "voteCount": 36,
    "commentCount": 1,
    "favoriteCount": 21
  },
  "publishedAt": "2024-02-15T06:46:33.000Z",
  "createdAt": "2024-02-15T06:46:33.000Z",
  "updatedAt": "2024-02-15T06:46:33.000Z",
  "topics": [
    {
      "name": "大模型",
      "id": "25402720",
      "url": "https://www.zhihu.com/topic/25402720"
    },
    {
      "name": "LLM",
      "id": "20660508",
      "url": "https://www.zhihu.com/topic/20660508"
    },
    {
      "name": "AI",
      "id": "19588023",
      "url": "https://www.zhihu.com/topic/19588023"
    }
  ],
  "column": {
    "id": "c_1741430207653949440",
    "name": "Hello，大模型",
    "url": "https://zhuanlan.zhihu.com/c_1741430207653949440"
  },
  "sourceGaps": [
    "fullText"
  ],
  "text": "<p data-pid=\"Aj57W3mu\">因工作原因，我从 2023 年中开始接触和学习大模型相关的知识。但由于过往的工作经历不在这个领域，在面对 AI、机器学习、深度神经网络、决策式 AI、生成式 AI、大模型、GPT 等一堆概念时，其实并不太清楚它们的逻辑关系，这可能不利于对大模型的进一步学习。我想有必要对 AI 技术有一个宏观上的了解，搞清楚大模型在 AI 技术栈中处于怎样的层次。这也是《Hello，大模型》的开篇文章，智识有限，不免会有些错误，希望不要给大家带来麻烦。</p><h2>1. 早期发展：符号人工智能（Symbolic AI）</h2><p data-pid=\"Hdhi_7Np\">人工智能（AI，Artificial Intelligence）作为计算机科学的一个分支，大概出现在 20 世纪 50 年代，核心目的是让计算机能够像人类一样思考、学习、推理、决策、交流，更好地解决复杂的问题，变得更加有用。</p><p data-pid=\"CvCbI9yd\">从 AI 的出现的 50 年代到 80 年代中前期，是 AI 发展的早期阶段，称为“符号人工智能（Symbolic AI）”。<b>“符号”可以简单的理解为语句和规则，</b>计算机的智能性主要依赖于：<b>领域专家和程序员将现实生活中的知识转换为规则并编程实现（一条简单的规则可能就是“if...els</b></p>",
  "summary": "<img src=\"https://pic4.zhimg.com/v2-e6ac24f0247c22fe48774666b5a980cd_200x112.png\" data-caption=\"样本、特征、数据集\" data-size=\"normal\" data-rawwidth=\"636\" data-rawheight=\"294\" data-watermark=\"watermark\" data-original-src=\"v2-e6ac24f0247c22fe48774666b5a980cd\" data-watermark-src=\"v2-f0b9322a4be16fcf6864840f5173ae32\" data-private-watermark-src=\"\" class=\"origin_image inline-img zh-lightbox-thumb\" data-original=\"https://pic4.zhimg.com/v2-e6ac24f0247c22fe48774666b5a980cd_r.png\"/>因工作原因，我从 2023 年中开始接触和学习大模型相关的知识。但由于过往的工作经历不在这个领域，在面对 AI、机器学习、深度神经网络、决策式 AI、生成式 AI、大模型、GPT 等一堆概念时，其实并不太清楚它们的逻辑关系，这可能不利于对大模型的进一步学习…"
}

Column row

Field	Type	What it does
`recordType`	string	Identifies the row as a `column`.
`source`	object	Holds the public column article page and its source identity.
`source.id`	string	Gives the article ID when available.
`source.url`	string URL	Gives the public article URL in the column.
`source.discoveryInput`	string	Shows the keyword, URL, handle, or ID that first found the listing.
`title`	string	Gives the listed article title.
`summary`	string	Gives the listed article excerpt when available.
`text`	string	Gives available public article text when the listing provides it.
`author`	object	Holds public article-author details.
`author.id`	string	Gives the author ID when available.
`author.name`	string	Gives the public author name.
`author.url`	string URL	Gives the public author profile URL.
`author.headline`	string	Gives the public author headline when available.
`author.isVerified`	boolean	Says whether Zhihu marks the author as verified.
`engagement`	object	Holds public engagement counts.
`engagement.voteCount`	integer	Gives the public vote count when available.
`engagement.likeCount`	integer	Gives the public like count when available.
`engagement.commentCount`	integer	Gives the public comment count when available.
`engagement.replyCount`	integer	Gives the public reply count when available.
`engagement.viewCount`	integer	Gives the public view count when available.
`engagement.favoriteCount`	integer	Gives the public favorite count when available.
`engagement.shareCount`	integer	Gives the public share count when available.
`publishedAt`	date-time string	Gives the UTC publication time when available.
`createdAt`	date-time string	Gives the UTC creation time when available.
`updatedAt`	date-time string	Gives the UTC update time when available.
`column`	object	Holds the public source column.
`column.id`	string	Gives the column ID when available.
`column.name`	string	Gives the column name when available.
`column.url`	string URL	Gives the column URL when available.
`topics`	array of objects	Lists public topics linked to the article listing.
`topics[].id`	string	Gives a topic ID when available.
`topics[].name`	string	Gives a topic name when available.
`topics[].url`	string URL	Gives a topic URL when available.
`sourceGaps`	array of strings	Lists public data areas not exposed for the row.

Genuine column row

This row is shortened. The JSON string "..." replaces the omitted text content.

{
  "recordType": "column",
  "source": {
    "url": "https://zhuanlan.zhihu.com/p/1962852227333990287",
    "id": "1962852227333990287",
    "discoveryInput": "https://zhuanlan.zhihu.com/column/c_1961542888350549465"
  },
  "title": "SOSP 2025 论文评述 Day 3 Session 13: OS Memory Management and Scalability",
  "column": {
    "id": "c_1961542888350549465",
    "url": "https://zhuanlan.zhihu.com/column/c_1961542888350549465"
  },
  "author": {
    "name": "IPADS-SYS",
    "id": "83b852754105538c6f7d4add05d8b39d",
    "url": "https://www.zhihu.com/people/ipadssys",
    "headline": "上海交通大学并行与分布式系统研究所官方知乎账号",
    "isVerified": false
  },
  "engagement": {
    "voteCount": 21,
    "commentCount": 0
  },
  "createdAt": "2025-10-18T10:20:48.000Z",
  "updatedAt": "2025-10-18T10:20:48.000Z",
  "summary": "本文是 SOSP 2025 Day 3 第十三个 session 的论文介绍，包含以下五篇论文： How to Copy Memory? Coordinated Asynchronous Copy as a First-Class OS Service CortenMM: Efficient Memory Management with Strong Correctness Guarantees Rearchitecting the Thread Model of In-Memory Key-Value Stores with μTPS FlexGuard: Fast Mutual Exclusion Independent of Subscription Scalable Address Spaces using Concurrent Inter…",
  "text": "..."
}

Video row

Field	Type	What it does
`recordType`	string	Identifies the row as a `video`.
`source`	object	Holds the public video page and its source identity.
`source.id`	string	Gives the video ID when available.
`source.url`	string URL	Gives the public Zhihu video URL.
`source.discoveryInput`	string	Shows the URL, keyword, handle, or ID that first found the video.
`title`	string	Gives the public video title.
`summary`	string	Gives the public video description or excerpt when available.
`text`	string	Gives public video text when available.
`author`	object	Holds public video-author details.
`author.id`	string	Gives the author ID when available.
`author.name`	string	Gives the public author name.
`author.url`	string URL	Gives the public author profile URL.
`author.headline`	string	Gives the public author headline when available.
`author.isVerified`	boolean	Says whether Zhihu marks the author as verified.
`engagement`	object	Holds public engagement counts.
`engagement.voteCount`	integer	Gives the public vote count when available.
`engagement.likeCount`	integer	Gives the public like count when available.
`engagement.commentCount`	integer	Gives the public comment count when available.
`engagement.replyCount`	integer	Gives the public reply count when available.
`engagement.viewCount`	integer	Gives the public view count when available.
`engagement.favoriteCount`	integer	Gives the public favorite count when available.
`engagement.shareCount`	integer	Gives the public share count when available.
`publishedAt`	date-time string	Gives the UTC publication time when available.
`createdAt`	date-time string	Gives the UTC creation time when available.
`updatedAt`	date-time string	Gives the UTC update time when available.
`topics`	array of objects	Lists public topics linked to the video.
`topics[].id`	string	Gives a topic ID when available.
`topics[].name`	string	Gives a topic name.
`topics[].url`	string URL	Gives a topic URL when available.
`media`	object	Holds public video metadata.
`media.videoUrl`	string URL	Gives a usable public video URL when available.
`media.coverImageUrl`	string URL	Gives the public cover image URL when available.
`media.durationSeconds`	number	Gives the video duration in seconds.
`media.width`	integer	Gives the video width in pixels.
`media.height`	integer	Gives the video height in pixels.
`sourceGaps`	array of strings	Lists public data areas not exposed for the row.

Genuine video row

{
  "recordType": "video",
  "source": {
    "url": "https://www.zhihu.com/zvideo/1689579617038585857",
    "id": "1689579617038585857",
    "discoveryInput": "https://www.zhihu.com/zvideo/1689579617038585857"
  },
  "title": "CMF课程 | 有没有更加科学有效的学习过程与方法？",
  "author": {
    "name": "涟一CMF创新研究",
    "id": "f28156af42ade98d4c366c81cd8bf64f",
    "url": "https://www.zhihu.com/people/cmf-59-52",
    "headline": "为机器人|家居家电|3C|汽车|美妆企业提供CMF咨询服务"
  },
  "engagement": {
    "voteCount": 2,
    "commentCount": 5,
    "favoriteCount": 3,
    "shareCount": 0,
    "viewCount": 186
  },
  "publishedAt": "2023-09-25T02:30:06.000Z",
  "updatedAt": "2023-09-25T02:34:23.000Z",
  "topics": [
    {
      "name": "CMF",
      "id": "19573320",
      "url": "https://www.zhihu.com/topic/19573320"
    },
    {
      "name": "CMF课程",
      "id": "28161640",
      "url": "https://www.zhihu.com/topic/28161640"
    },
    {
      "name": "CMF学习",
      "id": "28362175",
      "url": "https://www.zhihu.com/topic/28362175"
    }
  ],
  "media": {
    "videoUrl": "https://vdn3.vzuu.com/FHD/e3b432ac-569c-11ee-b122-96e157934a00-v8_f2_t1_7XXnYk3s.mp4?auth_key=1793744472-0-0-ddb6d214f6fc7e957a26702fe7eafe1e&bu=1513c7c2&c=avc.8.0&disable_local_cache=1&expiration=1793744472&f=mp4&pu=1513c7c2&v=tx",
    "coverImageUrl": "https://pica.zhimg.com/v2-d644514e3fb6b44a88c01d41d3bb5882_r.jpg?source=12a79843",
    "durationSeconds": 413.709,
    "width": 1920,
    "height": 1080
  },
  "summary": "涟一CMF课程分享：大家都知道这个行业是存在前景有一定发展的。但是如何去让自己做的收益更大？\n\n首先呢我们认为一个学习的金字塔，CMF的学习金字塔它分成三个阶段:第一个阶段是认知理解；第二个阶段是方法实践；第三个阶段是策略规划。\n\n什么意思呢？我们来看一下~\n\n想了解更多家居、家电、3C、汽车的CMF资讯、CMF课程、CMF报告、CMF设计咨询服务等内容，欢迎关注涟一CMF创新研究~"
}

Comment row

Field	Type	What it does
`recordType`	string	Identifies the row as a `comment`.
`source`	object	Holds the public comment page and its source identity.
`source.id`	string	Gives the comment ID when available.
`source.url`	string URL	Gives the public comment URL.
`source.discoveryInput`	string	Shows the content URL that first found the comment.
`text`	string	Gives the public comment text.
`author`	object	Holds public comment-author details.
`author.id`	string	Gives the author ID when available.
`author.name`	string	Gives the public author name.
`author.url`	string URL	Gives the public author profile URL.
`author.headline`	string	Gives the public author headline when available.
`author.isVerified`	boolean	Says whether Zhihu marks the author as verified.
`engagement`	object	Holds public comment engagement counts.
`engagement.voteCount`	integer	Gives the public vote count when available.
`engagement.likeCount`	integer	Gives the public like count when available.
`engagement.commentCount`	integer	Gives the public comment count when available.
`engagement.replyCount`	integer	Gives the public reply count when available.
`engagement.viewCount`	integer	Gives the public view count when available.
`engagement.favoriteCount`	integer	Gives the public favorite count when available.
`engagement.shareCount`	integer	Gives the public share count when available.
`createdAt`	date-time string	Gives the UTC comment creation time.
`content`	object	Holds the public content linked to the comment.
`content.recordType`	string	Identifies the linked content as a `question`, `answer`, `article`, or `video`.
`content.id`	string	Gives the linked content ID when available.
`content.title`	string	Gives the linked content title when available.
`content.url`	string URL	Gives the linked content URL.

Genuine comment row

{
  "recordType": "comment",
  "source": {
    "url": "https://www.zhihu.com/question/471506869/answer/1991187375/comment/4430293870",
    "id": "4430293870",
    "discoveryInput": "https://www.zhihu.com/question/471506869/answer/1991187375"
  },
  "text": "哈？[思考]",
  "author": {
    "name": "思悥",
    "id": "e2f3c8a858a97f324a950ec2e25a8e0a",
    "url": "https://www.zhihu.com/people/tian-cai-68-16",
    "headline": "小作坊磕盐民工 g-u-n.github.io"
  },
  "engagement": {
    "likeCount": 0,
    "replyCount": 1
  },
  "createdAt": "2021-07-11T03:08:52.000Z",
  "content": {
    "recordType": "answer",
    "id": "1991187375",
    "title": "请问Statistical learning 和 machine learning 具体有什么区别？",
    "url": "https://www.zhihu.com/question/471506869/answer/1991187375"
  }
}

Comment reply row

Field	Type	What it does
`recordType`	string	Identifies the row as a `commentReply`.
`source`	object	Holds the public reply page and its source identity.
`source.id`	string	Gives the reply ID when available.
`source.url`	string URL	Gives the public reply URL.
`source.discoveryInput`	string	Shows the comment URL that first found the reply.
`text`	string	Gives the public reply text.
`author`	object	Holds public reply-author details.
`author.id`	string	Gives the author ID when available.
`author.name`	string	Gives the public author name.
`author.url`	string URL	Gives the public author profile URL.
`author.headline`	string	Gives the public author headline when available.
`author.isVerified`	boolean	Says whether Zhihu marks the author as verified.
`engagement`	object	Holds public reply engagement counts.
`engagement.voteCount`	integer	Gives the public vote count when available.
`engagement.likeCount`	integer	Gives the public like count when available.
`engagement.commentCount`	integer	Gives the public comment count when available.
`engagement.replyCount`	integer	Gives the public reply count when available.
`engagement.viewCount`	integer	Gives the public view count when available.
`engagement.favoriteCount`	integer	Gives the public favorite count when available.
`engagement.shareCount`	integer	Gives the public share count when available.
`createdAt`	date-time string	Gives the UTC reply creation time.
`content`	object	Holds the public content linked to the reply.
`content.recordType`	string	Identifies the linked content as a `question`, `answer`, `article`, or `video`.
`content.id`	string	Gives the linked content ID when available.
`content.title`	string	Gives the linked content title when available.
`content.url`	string URL	Gives the linked content URL.
`parentComment`	object	Holds the parent comment relationship.
`parentComment.id`	string	Gives the parent comment ID when available.
`parentComment.url`	string URL	Gives the parent comment URL when available.
`parentComment.text`	string	Gives the parent comment text when available.

Genuine comment reply row

{
  "recordType": "commentReply",
  "source": {
    "url": "https://www.zhihu.com/question/471506869/answer/1991187375/comment/4429085608",
    "id": "4429085608",
    "discoveryInput": "https://www.zhihu.com/question/471506869/answer/1991187375/comment/4430293870"
  },
  "text": "[微笑]",
  "author": {
    "name": "麦契格斯",
    "id": "448b273d911b2d440cd603d834d3feeb",
    "url": "https://www.zhihu.com/people/matrix-58-77",
    "headline": "深度学习机器学习"
  },
  "engagement": {
    "likeCount": 0,
    "replyCount": 0
  },
  "createdAt": "2021-07-11T03:14:25.000Z",
  "content": {
    "recordType": "answer",
    "id": "1991187375",
    "title": "请问Statistical learning 和 machine learning 具体有什么区别？",
    "url": "https://www.zhihu.com/question/471506869/answer/1991187375"
  },
  "parentComment": {
    "id": "4430293870",
    "url": "https://www.zhihu.com/question/471506869/answer/1991187375/comment/4430293870",
    "text": "哈？[思考]"
  }
}

Person row

Field	Type	What it does
`recordType`	string	Identifies the row as a `person`.
`source`	object	Holds the public profile page and its source identity.
`source.id`	string	Gives the profile ID when available.
`source.url`	string URL	Gives the public profile URL.
`source.discoveryInput`	string	Shows the keyword, profile URL, handle, or ID that first found the profile.
`title`	string	Gives a public profile title when available.
`author`	object	Holds the public person identity used for the row.
`author.id`	string	Gives the author ID when available.
`author.name`	string	Gives the public display name.
`author.url`	string URL	Gives the public profile URL.
`author.headline`	string	Gives the public headline when available.
`author.isVerified`	boolean	Says whether Zhihu marks the person as verified.
`profile`	object	Holds public profile details and counts.
`profile.name`	string	Gives the public profile name.
`profile.username`	string	Gives the public username or handle when available.
`profile.bio`	string	Gives the public biography when available.
`profile.headline`	string	Gives the public profile headline when available.
`profile.avatarUrl`	string URL	Gives the public avatar URL when available.
`profile.isVerified`	boolean	Says whether Zhihu marks the person as verified.
`profile.verificationReason`	string	Gives the public verification reason when available.
`profile.followerCount`	integer	Gives the public follower count when available.
`profile.followingCount`	integer	Gives the public following count when available.
`profile.answerCount`	integer	Gives the public answer count when available.
`profile.articleCount`	integer	Gives the public article count when available.
`profile.questionCount`	integer	Gives the public question count when available.
`profile.videoCount`	integer	Gives the public video count when available.
`profile.location`	string	Gives the public location when available.
`sourceGaps`	array of strings	Lists public data areas not exposed for the row.

Genuine person row

{
  "recordType": "person",
  "source": {
    "url": "https://www.zhihu.com/people/zhang-jia-wei",
    "id": "f9de84865e3e8455a09af78bfe4d1da5",
    "discoveryInput": "https://www.zhihu.com/people/zhang-jia-wei"
  },
  "author": {
    "name": "张佳玮",
    "url": "https://www.zhihu.com/people/zhang-jia-wei"
  },
  "profile": {
    "name": "张佳玮",
    "username": "zhang-jia-wei",
    "headline": "公号：张佳玮写字的地方",
    "avatarUrl": "https://picx.zhimg.com/v2-d626957872f9b38fa594394db2d2e6e7_xl.jpg?source=32738c0c&needBackground=1",
    "isVerified": false,
    "followerCount": 3462551,
    "followingCount": 148,
    "answerCount": 6379,
    "articleCount": 1575,
    "questionCount": 1,
    "videoCount": 211
  },
  "sourceGaps": [
    "location"
  ]
}

💳 Pricing

This Actor uses pay-per-event pricing. The primary event is Content match. Each line below describes one public row type that can be emitted. The Store shows the current price for your account tier.

Buyer-facing event	Charge description
Content match	One public answer, question, article, video, or column returned by a keyword search or selected content type.
Hot question	One public hot-list question, with rank, heat, and top-answer details when shown.
Answer	One public answer returned for a supplied question, with available text and question context.
Article detail	One public article returned from a supplied article URL, with available text, author, topics, and dates.
Column article	One public article listing returned from a supplied public column URL.
Video	One public video returned from a supplied URL or selected content-type search.
Comment	One top-level public comment returned for a supplied content URL.
Comment reply	One public reply returned for a supplied comment, with its parent link when available.
Creator profile	One public creator profile returned from discovery or a supplied profile URL.

🔌 Integrations

Open the saved rows from the run's Storage tab or read the default dataset through the Apify API. The Actor's public data is ready for review and export in the dataset.

❓ FAQ

How does the Actor handle the same source found more than once?

It saves the first eligible occurrence immediately. If another keyword, URL, handle, or ID finds the same source item later, that later match is ignored and the saved row is not rewritten.

Can I run more than one keyword in one job?

No. Enter one keyword or phrase for keywordSearch. Start another run for a different search.

What happens when Zhihu does not expose a field?

The missing field is omitted from that row. sourceGaps can list fullText, author, engagement, publicationTime, topics, relatedContent, media, profileDetails, location, or verification.

Can I collect full answers and articles?

The Actor returns the available public body for answer and article retrieval. A source gap is recorded when Zhihu does not expose the full text.

Can I collect comments and comment replies separately?

Yes. Use contentUrls for top-level comments and commentUrls for replies. Reply rows keep the parent comment link when it is available.

What does an empty or 0 `maxItems` value do?

Both return all available results. Set a positive number when you want a smaller set to review.

No. It is for public Zhihu pages. Private, deleted, login-only, paywalled, and otherwise restricted content is outside the supported scope.

Which result type should I choose for a creator?

Choose creatorProfile for the public profile itself. Use contentTypes: ["people"] when finding people through keyword search.

Can I read the dataset with code?

Yes. Open the default dataset from the run or use the Apify API to read the saved rows.

What is Zhihu?

Zhihu (知乎) is a large Chinese question-and-answer site. People post questions, answers, articles, columns, and videos there. This Actor collects public Zhihu pages as structured rows.

📝 Changelog

v1.0 (03-11-2026)

Removed the creatorArticles choice from target. Zhihu no longer shows a creator's article list to signed-out visitors, so the Actor cannot collect it. There is no replacement. Change target to another value, such as creatorProfile, before you run saved inputs or tasks.
Removed the profile.lastActiveAt output field. Zhihu no longer shows this time to signed-out visitors. There is no replacement.
Keyword search with the default input works again and returns results.
The Actor now needs at least 1024 MB of memory.
Reliability improvements.

v0.0

Initial release.

🆘 Support

For issues, questions, or feature requests, file a ticket and I'll fix or implement it in less than 24h 🫡

Quora Search Scraper: Compare public Q&A pages with visible answers, topics, profiles, and counts.
Reddit Comments Search Scraper: Search public Reddit comments for another source of product and brand discussions.
Zhihu Scraper — Q&A, Answers, Articles, Columns: Use it for long-form Zhihu Q&A and article collection.
❓ Zhihu Search Scraper: Use it for public Zhihu search records from keywords, URLs, or IDs.
Zhihu Article Scraper: Use it for public Zhihu Zhuanlan article text and metadata.

Made with ❤️ by Maxime Dupré

Zhihu Scraper (知乎) | Search, Answers, Comments & Profiles

zen-studio/zhihu-scraper

Search Zhihu and collect answers, articles, comments, replies, creator profiles and trending questions. Export structured JSON, CSV or Excel.

Zen Studio

322

Zhihu Scraper — Q&A, Answers, Articles, Columns

sian.agency/zhihu-scraper

Zhihu scraper — extract long-form Mandarin Q&A, expert answers, articles & column posts. Keyword search, question answer threads, article detail, column article list. China market research, LLM training data, competitive intel. Four operations, one clean dataset per run. No API key.

SIÁN OÜ

264

YouTube Transcript Scraper

seemuapps/youtube-transcript-scraper

Extract full transcripts and timestamped captions from any YouTube video in your chosen language and export them as JSON or CSV.

Seemu Scraping

359

5.0

(1)

Bulk Domain Rating Checker

datascraperes/bulk-domain-rating-checker

Check Domain Rating by Ahrefs for up to 10,000 domains or URLs with clean three-field results, resilient retries, and transparent per-result pricing.

DataScraperES

231

Keyword Research Tool: Search Volume, CPC & Keyword Ideas

santhej/keyword-research-tool

Bulk keyword research: accurate Google search volume, CPC, competition & thousands of keyword ideas for any seed or market. Clean JSON/CSV for SEO, PPC & content. Up to 5,000 keywords/run, 190+ countries. No subscriptions, no API keys.

Santhej Kallada

403

5.0

(1)

Keyword Search Volume, CPC & Difficulty — Google + Bing Bulk

memo23/keyword-metrics

Keyword search volume, CPC, competition, difficulty and search intent from Google Ads and Bing, with up to ~7 years of monthly history and a trend flag. One flat fee covers up to 1000 keywords per run. Expand seeds into keyword ideas or pull a competitor's ranked keywords. 111 countries.

Muhamed Didovic

988

Reddit Comments Search Scraper

maximedupre/reddit-comments-search-scraper

Search public Reddit comments by keyword, with optional subreddit filters. Export comment text, authors, scores, links, post context, timestamps, and search details.

Maxime Dupré

332

Douyin Analytics Scraper

maximedupre/douyin-analytics-scraper

Collect current public Douyin hot-search topics. Export ranks, hot values, video counts, source IDs, cover image URLs, source timestamps, and scrape times to an Apify dataset.

Maxime Dupré

DataForSEO API Gateway - Run Any Endpoint, No Subscription

santhej/dataforseo-api-gateway

Call ANY DataForSEO API endpoint pay-per-use through Apify — no DataForSEO account or subscription required. SERP, Keywords, Labs, Backlinks, On-Page, Merchant, Business Data & more. Just pass the endpoint path + payload and get structured JSON back.

Santhej Kallada

539

5.0

(1)

Twitter Organization Affiliates Scraper

maximedupre/twitter-organization-affiliates-scraper

Export public X/Twitter organization affiliates from URLs, @handles, /affiliates paths, or numeric IDs. Get handles, names, bios, follower counts, verification flags, business labels, badge URLs, profile links, and scrape timestamps.

Maxime Dupré

Zhihu Scraper (知乎)

🔎 Find public Zhihu research records

📚 Explore Zhihu research records

▶️ Choose a Zhihu collection

⚙️ Input

🧾 Output

💳 Pricing

🔌 Integrations

❓ FAQ

How does the Actor handle the same source found more than once?

Can I run more than one keyword in one job?

What happens when Zhihu does not expose a field?

Can I collect full answers and articles?

Can I collect comments and comment replies separately?

What does an empty or 0 maxItems value do?

Does the Actor collect private or login-only pages?

Which result type should I choose for a creator?

Can I read the dataset with code?

What is Zhihu?

📝 Changelog

🆘 Support

🔗 Related Actors

You might also like

Zhihu Scraper (知乎) | Search, Answers, Comments & Profiles

Zhihu Scraper — Q&A, Answers, Articles, Columns

YouTube Transcript Scraper

Bulk Domain Rating Checker

Keyword Research Tool: Search Volume, CPC & Keyword Ideas

Keyword Search Volume, CPC & Difficulty — Google + Bing Bulk

Reddit Comments Search Scraper

Douyin Analytics Scraper

DataForSEO API Gateway - Run Any Endpoint, No Subscription

Twitter Organization Affiliates Scraper

What does an empty or 0 `maxItems` value do?