Weibo Scraper: Posts and Profiles by User ID

Weibo is the main Chinese microblogging platform, and for brand monitoring, KOL tracking and competitor watching in the China market it is where the public conversation happens. The problem is that almost every scraping guide for it is wrong in the same way, and the failure mode misleads you.

Weibo blocks non China IP addresses at the connection level. Not with a 403, not with a captcha, and not with a login wall. The connection simply times out before Weibo’s application layer ever sees the request. So a script that works perfectly in principle produces timeouts, and the natural conclusion is that your parser is broken or that you have been detected as a bot. Neither is true. You are just not in China.

Try it live: Weibo Scraper: Posts and Profiles by User ID. Pay only for results delivered, no result no charge.

The Weibo Scraper handles the routing as part of the run, so the interesting part is left to you.

What you get

Two row types, distinguished by record_type.

post-scraped, one per post:

Field	Notes
post_id	Stable post identifier
text	Post body
published_at	Publication timestamp
source_client	The device or client string Weibo attaches, for example iPhone or Android
reposts_count, comments_count, likes_count	The three engagement metrics Weibo exposes
image_urls	Attached images, as an array

profile-scraped, optional, one per user: display name, follower count, verification status and bio. Off by default because it costs one extra request per user, and most monitoring runs only need the posts.

Finding a user ID

Open the profile in a browser. The URL contains the numeric ID:

https://weibo.com/u/1699432410
                    ^^^^^^^^^^ this is what you pass

Store the numeric ID rather than the handle. Display names on Weibo change, the numeric ID does not.

Tracking engagement over time

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

run = client.actor("themineworks/weibo-scraper").call(run_input={
    "profileIds": ["1699432410"],
    "maxPostsPerProfile": 50,
    "includeProfileInfo": True,
    "proxyConfiguration": {
        "useApifyProxy": True,
        "apifyProxyGroups": ["RESIDENTIAL"],
        "apifyProxyCountry": "CN",
    },
})

posts = []
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
    if row.get("record_type") == "profile-scraped":
        print(f"{row.get('screen_name')}: {row.get('followers_count'):,} followers")
    elif row.get("record_type") == "post-scraped":
        posts.append(row)

posts.sort(key=lambda p: p.get("likes_count") or 0, reverse=True)
for p in posts[:5]:
    print(f"{p['likes_count']:>7} likes  {p['text'][:70]}")

Run this on a schedule and store by post_id. Because engagement counts move after publication, re-reading the same posts over several days tells you how a post actually performed rather than how it started.

Cost reality

This is the one actor in this batch that needs a paid proxy tier, because China residential routing is not optional. That makes it more expensive to run than a no proxy actor, and the pricing reflects it. It is still far cheaper than a browser based approach, because the actor talks to JSON endpoints rather than rendering pages.

If a run returns nothing, check the proxy first. A timeout with zero posts and no error is the signature of a request that never reached China, not of a parsing problem.

Pricing

Pay per result delivered, from $2.25 per 1,000 on higher Apify plans and $3.00 per 1,000 on the free plan. Posts and profile rows are priced identically. A profile that yields nothing is not charged.

Use it from an AI agent

https://mcp.apify.com/?tools=themineworks/weibo-scraper

Related

• Bilibili Scraper for Chinese long form video and creator data
• Threads Scraper for the equivalent western microblog conversation
• Twitter/X Scraper for cross platform brand monitoring