For AI

Answers

Is the heat number on a hot search list a count of searches?

Short answer

No. On Weibo the heat number is a composite of searches, posts, reads and interaction; on Douyin it is the platform's own index; Xiaohongshu's score is what the app labels as people viewing; Bilibili's popular board is ranked videos whose heat is their view count. okdata returns each as the heat field of its hot list, $0.002 for the Weibo board, and the scales are not comparable across platforms.

A topic sits at the top of Weibo with 878,353 next to it, and a report quotes that as "878,353 searches". It is not. None of the four Chinese hot lists publishes a raw search count. Each prints a number of its own making, and the four numbers measure different things on different scales.

What the number is on each platform

Platform What heat is Range on a board we fetched on 11 November 2026
Weibo hot search Weibo's heat value. Weibo's 2021 rules give it as (search + discussion + spread) × interaction rate: searches, original posts and reposts, reads, and interaction on the results page about 72,000 at #50 to 878,000 at #1
Douyin hot search Douyin's hot value. We could not find a public definition of how it is calculated about 7.7 million to 11.6 million; the pinned first row has none
Xiaohongshu hot topics The score the app shows as "N万人在看", people viewing the topic about 18 million to 24 million
Bilibili popular Not a search list: each row is a video and heat is its view count about 138,000 to 2.6 million on the first page

Two more details change how you read rank. Since April 2026 Weibo also ranks on public reach, attention within a field and content quality, so a row can sit above one with more heat. And Weibo's 2021 rules describe two advertising slots below positions 3 and 6, marked 商; okdata removes promoted rows from the Weibo board.

How to use heat without misreading it

  • Compare within one board. Heat tells you how far ahead the top topic is of the tenth on the same platform at the same moment.
  • Compare over time on one platform. Store every snapshot with meta.fetched_at and you can see a topic climb, peak and fade.
  • Across platforms, compare rank or share of the board's top row, never raw heat. 878,000 on Weibo and 11 million on Douyin do not mean Douyin cares twelve times more.
  • Do not call it searches in a report. "Weibo heat value" or "Douyin hot value" is accurate.
Snapshot Calls Cost
All four boards once 4 $0.026
Weibo every hour for a day 24 $0.048
All four, hourly for 30 days 2,880 $18.72

Example

Fetch the four boards and express each row's heat as a share of its own board's top row, which is comparable:

import os, requests

H = {"Authorization": f"Bearer {os.environ['OKDATA_API_KEY']}"}
BOARDS = {
    "weibo": "https://api.okdata.online/v1/weibo/trends/list",
    "douyin": "https://api.okdata.online/v1/douyin/trends/list",
    "xiaohongshu": "https://api.okdata.online/v1/xiaohongshu/trends/list",
    "bilibili": "https://api.okdata.online/v1/bilibili/trends/list",
}

for name, url in BOARDS.items():
    body = requests.get(url, headers=H, params={"fields": "rank,name,heat"}, timeout=60).json()
    if body["code"] != "ok":
        print(name, body["code"]); continue
    rows = [r for r in body["data"]["items"] if r.get("heat")]
    top = max(r["heat"] for r in rows)
    for r in rows[:5]:
        print(f"{name:<12} #{r['rank']:<3} {r['heat'] / top:6.1%}  {r['name']}")

Limits

  • heat can be null, for example on Douyin's pinned row. Keep the row; its rank still counts.
  • The platforms change their formulas without notice, as Weibo did in 2026. Long time series should note the date of any known change.
  • Rank comes from the board's own order. The hot list tools show the same boards in a browser, and the Weibo hot search endpoint page lists every field.

To keep a history of any board, see building a Douyin hot search history and one API for China's hot lists.

Sources

  1. IT之家: Weibo publishes its hot search rules and heat formula (2021-08-23)ithome.com · read
  2. IT之家: Weibo hot search ranking adds three dimensions (2026-04-29)ithome.com · read

Where people asked this (read 2026-11-12)

Updated

Connect once, then just ask.

Every new key comes with $0.10 to try it. One sentence connects your agent.

Start free
Connect your AI

Paste this into your agent

Read https://okdata.online/agents.md and set up okdata for me; my key is in the OKDATA_API_KEY environment variable.

Claude Code, Cursor, Codex and other agents that can open a link set themselves up from it.

MCP serverhttps://api.okdata.online/mcp