Перейти к навигации

How do I continue from a shared Zhihu answer to other viewpoints and comments?

A Zhihu answer link contains both a question ID and an answer ID. Pass the answer ID to zhihu.web.fetch_answer_detail to read the shared answer, or pass the question ID to zhihu.web.fetch_question_answers to compare the other answers. Save both IDs separately, and add the comment ID when you read a discussion.

For example, a colleague shares an answer about daytime home-projector use. Record the answer’s conditions before comparing other answers under the same question: room light, screen and model may explain why conclusions differ.

The numbers in the table are format examples:

Link shapeIdentifier to preserveTask
zhihu.com/question/123question_id: "123"Question details and its answers
zhihu.com/question/123/answer/456Question123 and answer456 separatelyOne answer, or other answers under question123
zhuanlan.zhihu.com/p/789article_id: "789"Column-article details
zhihu.com/people/<token>Profile url_tokenUser information/content using the action’s declared input

Keep identifiers as original strings. When a full link already shows the IDs, take them from the path; no extra search is needed. A column article and an answer are different objects, even when both discuss the same subject, so archive them separately.

Read the shared answer first

Pass the answer_id to zhihu.web.fetch_answer_detail:

{
"action": "zhihu.web.fetch_answer_detail",
"params": {"answer_id": "<confirmed_answer_id>"}
}

Send the JSON to POST https://api.beatapi.io/v1/social-data/call with normal BeatAPI Bearer authentication. Reuse complete saved material first. In the returned answer, check the author and the parent question, and save the source and the collection time.

Compare other responses to the question

Pass the question ID to zhihu.web.fetch_question_answers:

{
"action": "zhihu.web.fetch_question_answers",
"params": {"question_id": "<confirmed_question_id>", "limit": 5, "offset": 0, "order": "default", "cursor": "", "session_id": ""}
}

Read a first batch, select the answers relevant to daytime conditions, and fetch individual details only where needed. Record each conclusion with its context; a vote count does not show that everyone agrees. See the question-research guide.

For the example link above, record “question 123 › answer 456.” When you open the answer list for question 123, record a separate answer ID for each answer you add. If you only have a short link, open the short link to get the full URL, then take the IDs from the full path.

Which object identifies comments and replies?

Answer comments take the answer_id in zhihu.web.fetch_comment_v5; replies take the comment_id in zhihu.web.fetch_sub_comment_v5. For example, to add context to a comment that corrects a model, request the replies of that comment; do not pass the question ID as the comment target.

In the comment actions, order_by is score for popular or ts for newest. The answer list uses its own order values, default and updated. Record a question’s answer count, an answer’s comment count and the number of items you actually read as three separate figures.

What is the user-identifier exception?

A profile url_token is not a numeric ID that every action accepts. In particular, zhihu.web.fetch_user_segments names its input user_url_token but requires the member hash id. Before calling zhihu.web.fetch_user_segments, confirm the member hash id; do not reuse a profile token because the field names look alike.

Prompt for the assistant:

Read this projector answer, keeping question ID, answer ID, author and original source. Review an initial batch of other answers and compare their environments and conditions. Preserve parent context for comment corrections and attach specific sources to conclusions.

Budget the selected answer pages, details and comments with current pricing.

Answer lists versus search · Question-answer research