> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.beatapi.io/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.beatapi.io/_mcp/server.

# How do I finish Zhihu research after a 40362 response or signing problem?

> Resume from saved questions and answers, fill specific text or comment gaps, and retain conditions, sources and unfinished scope.

**Keep the viewpoints and sources you already saved, then fill each gap by question ID or answer ID through BeatAPI, for example with `zhihu.web.fetch_answer_detail` for an answer that has only an excerpt.** An interruption does not require restarting the project from keyword search.

For example, in research on daytime projector use, three questions and some of their answers are already saved when the old collector stops. Sort the material into complete originals, answers missing context and comment threads needing expansion, then choose the next read.

## Inventory the material

| Existing material                             | Next step                                               |
| --------------------------------------------- | ------------------------------------------------------- |
| Complete answer with its source               | Continue comparing claims and conditions                |
| Verified question ID, more experiences needed | Read that question's answers and preserve the new batch |
| Verified answer ID, excerpt only              | Retrieve its detail and reconnect it to the question    |
| A possible comment correction                 | Confirm objects and add relevant context                |
| Only a keyword or unclear short link          | Resolve identity before adding it to confirmed material |

Write each gap as an object and a reason, such as “answer A needs context” or “question B lacks a bright-room experience”; a list in that form is easier to resume than “Zhihu is not finished.”

## Retrieve an answer already identified

Call `zhihu.web.fetch_answer_detail`:

```json
{
  "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. Choose answer lists, details or comments according to the missing material; see [question research](/guides/zhihu-question-archive) and [comment context](/faq/zhihu-comment-context).

The old collector's pagination state does not carry over to `zhihu.web.fetch_question_answers`. To continue a question's answer list, save the actual `cursor`, `session_id` and request conditions returned by `zhihu.web.fetch_question_answers`. If no usable state is available, start again with a bounded first batch and match the actual answer IDs against the saved originals.

## How should the old error be interpreted?

First identify which request failed: the direct Zhihu request, the old collector or the BeatAPI call. Keep the target, the timestamp and the response, so you can find where the work stopped.

| What you see                                              | What to do next                                                                                                                            |
| --------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------ |
| Zhihu returns `40362` with a temporary access restriction | Save the message and mark the page as a failed read                                                                                        |
| The old code uses `x-zse-96`                              | Leave the header with the old web request. To read a known answer through BeatAPI, use the answer ID and Bearer authentication shown above |
| BeatAPI returns an authentication error                   | Check the request URL and Bearer header                                                                                                    |
| The collector stopped before saving a response            | Find the last readable page in your archive and put the remaining targets on the follow-up list                                            |

## How does new material update previous findings?

1. **Link old and new.** Associate the material by question ID and answer ID, and keep the old and new originals as separate files.
2. **Label the new material.** Record each collection time and whether the new material is a repeated reading, added context, an author edit or a comment correction.
3. **Recheck earlier summaries.** Recheck every summary that was based on an excerpt once you have more context. Where earlier differences in opinion come from different environments, add those conditions back to the research table.
4. **Cite the change.** For each finding that the new material supports or overturns, attach the wording and the source.

For example, you saved three answers to question 123 and quoted answer 456 from an excerpt only. Read answer 456 in full first. If the full text says the blackout curtains were closed, add that condition to the finding and cite the passage. Reuse the other two archived answers, and leave failed pages on the follow-up list.

Prompt for the assistant:

> Inventory saved Zhihu originals and list gaps by question and answer. Continue analyzing existing material, retrieving only missing text or context. Preserve conditions and sources, explain what new material changes, and separate unresolved identities, unread objects and failed reads.

Hand off the supported findings, a condition comparison and the unfinished items. Budget only the gaps with current [pricing](https://beatapi.io/pricing). Configure storage, scheduling and later notifications in your own task environment.