Baidu Tieba Scraper 百度贴吧 - Forum Threads, Replies, Sub-replies & Search

Scrape Baidu Tieba (百度贴吧), China's largest forum platform, and get clean JSON. One Actor covers everything:

• Forum threads (吧内帖子): the thread list of any forum (吧), sorted by latest reply or newest, optionally only featured (精品) threads, plus forum stats (members, posts, threads).
• Replies (楼层): every floor of a thread with full text, images, videos, author and like counts.
• Sub-replies (楼中楼): the nested replies under each floor.
• Keyword search (全站搜索): find threads across all of Tieba by keyword, newest first.

Key points:

• No login, no cookies. Only public data is collected, and your account is never involved.
• $2 per 1,000 results, one flat price for threads, replies and sub-replies.
• Fails loudly. If Baidu blocks the requests, the run fails with a clear message. It never silently returns an empty dataset.

Input

Field	Description
mode	forum, thread or search
forumNames	(forum) Forum names, with or without 吧, e.g. ["电动车", "李毅吧"]
threadUrlsOrIds	(thread) https://tieba.baidu.com/p/11055481323 or 11055481323. Replies are always included in this mode
searchKeywords	(search) Keywords, e.g. ["电动车 电池"]
maxThreadsPerSource	Threads per forum or keyword (default 50)
onlyGood	(forum) Only featured threads (精品贴)
sortBy	(forum) reply (latest reply) or create (newest thread)
includeReplies, maxRepliesPerThread	Add the replies (floors) of each thread (default 100 per thread)
includeSubReplies, maxSubRepliesPerFloor	Add the sub-replies under each floor (default 20 per floor)
requestDelaySecs	Pause between requests, minimum 1.5 s

{
  "mode": "forum",
  "forumNames": ["电动车"],
  "maxThreadsPerSource": 100,
  "sortBy": "create",
  "includeReplies": true,
  "maxRepliesPerThread": 50
}

Output

Every dataset item has a type: forum, thread, reply, sub_reply or error. They all share these fields: platform, tid, pid, forumName, forumId, title, text (HTML stripped, emoticons as [name]), media (images, videos and links), authorId, authorName, authorNickname, authorPortrait, createdAt (ISO 8601, UTC), url and scrapedAt.

{
  "platform": "tieba",
  "type": "thread",
  "tid": 10200085026,
  "forumName": "电动车",
  "forumId": 2767,
  "title": "我好像被骗了，还有没被骗的",
  "text": "现在买完了qq不理人，报案有用吗",
  "media": [{"type": "image", "url": "http://tiebapic.baidu.com/forum/pic/item/7c11b05....jpg"}],
  "authorId": 879898799,
  "authorName": "墨轩七月",
  "authorNickname": "墨轩七月",
  "createdAt": "2025-11-09T11:03:57+00:00",
  "url": "https://tieba.baidu.com/p/10200085026",
  "replyCount": 51,
  "agreeCount": 47,
  "lastReplyAt": "2026-08-05T01:37:42+00:00",
  "isTop": true
}

{"type": "reply", "tid": 10200085026, "pid": 152837796035, "floor": 2, "text": "叫巅峰车行的，还在吧里发帖", "authorNickname": "墨轩七月", "agreeCount": 0, "disagreeCount": 0, "subReplyCount": 0, "createdAt": "2025-11-09T11:04:57+00:00"}

{"type": "sub_reply", "tid": 10200085026, "pid": 152837843626, "parentPid": 152837830934, "text": "是啊蠢了", "authorNickname": "墨轩七月", "createdAt": "2025-11-09T11:18:14+00:00"}

{"type": "forum", "forumName": "电动车", "forumId": 2767, "memberCount": 359231, "postCount": 3243414, "threadCount": 127018, "category": "生活/交通工具"}

Field details:

• Thread items also have replyCount, agreeCount, lastReplyAt, isTop, isGood and abstract. In forum mode text is Tieba's short abstract. When includeReplies is on, text is replaced by the full first post.
• Reply items also have floor, agreeCount, disagreeCount and subReplyCount. Floor 1 (the opening post) is inside the thread item, so replies start at floor 2.
• Sub-reply items also have parentPid.

Use cases

• Consumer and brand research in China: what real users say about a product, brand, car or EV model, with reply and like counts.
• Sentiment and topic analysis: full reply threads with nested discussions, ready for an LLM or NLP pipeline.
• Lead generation and community monitoring: track new threads in niche forums (second-hand trading, local city forums, product forums).

Use it from code, n8n or an AI agent

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("xiaoliang527/tieba-posts-comments-scraper").call(run_input={
    "mode": "search", "searchKeywords": ["电动车 电池"], "maxThreadsPerSource": 40})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    if item["type"] == "thread":
        print(item["forumName"], item["title"], item["replyCount"])

MCP (Claude, Cursor): https://mcp.apify.com?tools=xiaoliang527/tieba-posts-comments-scraper with your Apify token as Bearer header.

Pricing

Event	Price
Result (thread, reply or sub-reply)	$0.002 ($2 / 1,000)
Forum info item	free
Errors (forum not found, deleted thread, invalid URL)	free

Set a maximum cost per run in the run options and the Actor stops cleanly when it is reached.

Limits

• Only public content. Deleted, hidden or login-only content is not available.
• Search returns about 20 threads per page, newest first. Very deep result pages may not be available.
• In search results some authors have only a nickname (authorName is null).
• Sub-replies need one extra request for each floor that has many of them, so runs with includeSubReplies are slower.
• Requests are sent one at a time with at least 1.5 s between them, to be gentle with Tieba. About 30 threads, 30 replies or 20 search results come back per request.
• Very large forums report up to 1,000 pages of threads. Use maxThreadsPerSource to limit how deep the Actor goes.
• If Baidu starts showing its security check (百度安全验证), the run stops with an error instead of returning partial or empty data silently. Try again later.