> 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 read more Weibo search pages without counting repeated posts as new feedback?

> Keep the query consistent, distinguish repeated content and changed observations, and deliver feedback with actual coverage and stopping reasons.

For page two of a Weibo search, call `weibo.web_v2.fetch_realtime_search` again with the original `query` and `page:2`; for a different search term, start again at page one. Save each page's response and query time, then deduplicate posts by the actual post ID or a verified original link, so that a post seen on two pages counts once.

For example, page one mostly contains launch announcements and page two surfaces a call-quality report from the subway: add the subway report to the findings. If the same report also appeared on page one, keep both page references and count the report once.

## How do you continue?

`weibo.web_v2.fetch_realtime_search` numbers `page` from 1. After saving page one, request page 2 as below, with the placeholder replaced by the exact page-one query:

```json
{
  "action": "weibo.web_v2.fetch_realtime_search",
  "params": {"query": "<same_query_as_page_1>", "page": 2}
}
```

Send the JSON to `POST https://api.beatapi.io/v1/social-data/call` with the same authentication as page one. `weibo.web_v2.fetch_realtime_search` has no `cursor` input, and these steps do not set up continuous collection. When you change the query, start a new observation at page one.

## How do you combine pages for reading?

Keep each page's actual inputs, collection time, request ID and original response. Match posts across pages by the stable identifier or original link in the response. `weibo.web_v2.fetch_realtime_search` has no recorded fixed list or identifier fields, so inspect the saved output instead of copying another platform's field paths.

| Situation                               | How to record it                                                                 |
| --------------------------------------- | -------------------------------------------------------------------------------- |
| One post appears on two pages           | Keep one content object and retain both page/time observations as sources        |
| One post's text or metrics changes      | Preserve both observations for review rather than replacing the earlier material |
| Many accounts repeat the same statement | Keep sources and mark repeated retellings during topic analysis                  |
| Identity cannot be established          | Leave it for review rather than merging different authors by matching text alone |

After grouping, check whether an issue appears across different authors and usage conditions. Repetition points to a lead only after you separate page overlap, reposts and independent experiences.

## When should you stop adding pages?

Set a fixed page budget for the round. After each page, check whether the page added useful material, then choose one action:

* **Enough excerpts and context support the feedback list:** stop paging and process the questions.
* **The page is empty or mostly repeats:** save the page's outcome, then end the round or revise the research question.
* **A concrete question still needs material and the budget allows:** request the next page.
* **The page fails or its structure cannot be interpreted:** save the incomplete record and resume from the saved state later.

An empty page describes one observation, not the absence of feedback across Weibo. Dynamic search results can move between requests, so record the range you actually read and revisit later if needed.

## What coverage statement should the team receive?

For example: **“We saved two pages each for the brand and model queries, matched repeated posts, and organized connection excerpts and missing usage conditions. Items without publication times are separate; comments were not read.”** Use the counts you actually recorded; do not estimate post volume from the number of pages.

Prompt for the assistant:

> Continue the same query on the next page, retaining original results and time. Match overlap with prior pages and explain the useful conditions or questions added. Organize feedback when it is sufficient; explain what remains missing before expanding. Report actual pages, duplicate handling and unfinished work.

Each added page is one more search request in the plan, priced at current [pricing](https://beatapi.io/pricing). Rereading and organizing saved pages does not need a new search.

[Brand terms to feedback](/faq/weibo-keyword-search) · [Trend research](/faq/weibo-hot-search)