Bilibili Scraper

Turn Bilibili into structured Gen-Z reaction data. Pull videos, real danmaku (弹幕) text, comment threads, creator profiles, and live-room status from China's biggest youth video platform. No login, no API key, no VPN. In seconds, not hours.

How it works

✨ Why teams choose this over other Bilibili scrapers

Paying for a "danmaku" scraper that only gives you a number? Searching English brand names and getting almost nothing back? Stuck looking up creators one ID at a time?

• 🎯 Get the actual danmaku text, not just a count. Most Bilibili scrapers report a danmaku number and stop there. This one pulls the real scrolling-comment text, thousands per video, with the video timestamp, display style, and a per-user hash so you can chart reactions second by second and filter spam.
• 🧩 Multi-part videos, fully covered. Bilibili tutorials and compilations often have dozens of parts (分P). Danmaku collection walks every part, not just the first one, so a 100-episode course gives you the whole reaction stream.
• 💬 Comment threads, not just top comments. Every comment comes back with its reply samples linked by parentCommentId, which roughly triples the text you get per video.
• 📺 Live status for 50 creators in one call. Point the live mode at a list of creator IDs and get who is live right now, stream titles, categories, and live viewer counts. Built for scheduled KOL monitoring.
• 🔎 Find creators by keyword, not just by ID. Search Bilibili's creators by niche and get their follower counts, levels, and video totals, so you can build a KOL shortlist instead of hunting profile links by hand.
• 📚 Complete creator catalogs with full metrics. Walk a creator's entire upload history, with views, likes, coins, favorites, shares, replies, and danmaku on every row, where most scrapers flake out or return partial stats.
• 📇 Outreach handles, extracted for you. When a creator lists a WeChat, Weibo, QQ, or business email in their bio, it comes back as a clean contacts object, so your KOL shortlist is ready to contact, not just to read.
• 🌐 Search English, get Chinese results. Bilibili indexes by Chinese text, so "耐克" returns far more than "Nike". Auto-localize searches both and merges them, so you get full native recall even when you search in English.
• 📊 Every Bilibili metric, structured. Views, likes, coins (投币), favorites (收藏), danmaku, shares, and replies on every video, plus tags and categories, ready for a spreadsheet or a model.
• 🧠 Tag reactions with sentiment, built in. Turn on sentiment analysis and every video, comment, and danmaku comes back scored positive, neutral, or negative, so you can chart how a topic is received without a second tool.
• 🔁 Turn any search into a daily monitor. Delta mode returns only videos you have not seen before, so a schedule becomes a hands-off "what's new" feed and quiet days cost almost nothing.
• 🚀 No login, no setup, fast. Works on public data with zero authentication, and fetches videos in parallel. Point it at a keyword or a URL and go.
• 💸 Pay for what each row is worth. Videos, comments, danmaku, and live checks are billed as separate result types, so a 2,000-danmaku pull does not cost like 2,000 videos.

🎯 What you can do with Bilibili data

Team	What they build
AI / LLM training teams	Chinese danmaku and comment corpora for sentiment models and Gen-Z slang classifiers, pulled at scale on a schedule
Equity research desks	Gen-Z reaction sentiment on consumer brands (Pop Mart, BYD, gaming and anime IPs) as a leading indicator
Brand monitoring agencies	Track Western-brand mentions, viral reactions, and creator partnerships across Chinese youth culture
Gaming and entertainment	Monitor Chinese reception of game launches and anime adaptations on the community that drives them
KOL and influencer agencies	Discover Bilibili creators by niche and vet them by coins, favorites, and followers before you reach out
Academic NLP researchers	Study danmaku behavior, Gen-Z language trends, and Chinese digital culture with clean structured data

📥 Input parameters

Parameter	Description
mode	search / search_users / video_detail / video_comments / video_danmaku / user_videos / live_info / popular
searchQuery	Keyword, in Chinese or English (search modes)
autoLocalize	Also search a brand's Chinese name and merge results (on by default)
searchAliases	Extra term variants to search and merge
sortOrder	Relevance, most views, newest, most danmaku, most favorites, most comments
durationFilter	Filter search by video length
pubtimeBegin / pubtimeEnd	Publish-date range (YYYY-MM-DD)
videoUrls	Video URLs or BV IDs (detail / comments / danmaku modes)
userIds	Creator IDs (mid) for creator videos and live status
category	Category filter for trending
maxItems	Total results cap
maxComments / maxDanmaku	Per-video caps
deltaMode / deltaStateKey	Recurring "only new videos" monitoring
sentimentAnalysis	Tag videos, comments, and danmaku with Chinese sentiment

📊 Output data

Video:

{
    "type": "video",
    "bvid": "BV1rz7569EWw",
    "title": "Example video title",
    "url": "https://www.bilibili.com/video/BV1rz7569EWw",
    "viewCount": 13518564,
    "likeCount": 1311984,
    "coinCount": 750521,
    "favoriteCount": 371694,
    "danmakuCount": 23231,
    "replyCount": 21942,
    "authorName": "Creator Name",
    "authorMid": 5294454,
    "publishDate": "2026-06-08T12:00:00Z",
    "category": "Life",
    "tags": ["review", "life"],
    "durationFormatted": "7:19"
}

Danmaku:

{
    "type": "danmaku",
    "videoBvid": "BV1rz7569EWw",
    "page": 1,
    "text": "肉眼可见的清纯",
    "progressMs": 417,
    "timestamp": 1775822400,
    "displayMode": 1,
    "fontSize": 25,
    "userHash": "ccae0794"
}

Live status:

{
    "type": "live",
    "mid": 50329118,
    "name": "哔哩哔哩英雄联盟赛事",
    "liveStatusLabel": "live",
    "title": "HLE vs LYON",
    "online": 607494,
    "areaName": "游戏赛事",
    "liveUrl": "https://live.bilibili.com/7734200"
}

Creator:

{
    "type": "user",
    "mid": 546195,
    "name": "老番茄",
    "fans": 20477550,
    "level": 6,
    "archiveCount": 664,
    "profileUrl": "https://space.bilibili.com/546195"
}

🗂️ Data fields

• Video: bvid, aid, cid, title, description, url, thumbnailUrl, duration, durationFormatted, viewCount, likeCount, coinCount, favoriteCount, shareCount, danmakuCount, replyCount, authorName, authorMid, publishDate, publishTimestamp, category, tags, partCount, videoParts, scrapedAt
• Comment: commentId, parentCommentId (for sub-replies), text, likeCount, replyCount, createdAt, authorName, authorMid, authorAvatar, authorLevel, videoBvid, scrapedAt
• Danmaku: videoBvid, cid, page, text, progressMs, timestamp, color, displayMode, fontSize, userHash, scrapedAt
• Creator: mid, name, face, sign, level, fans, archiveCount, profileUrl, contacts (wechat / weibo / qq / email when present in the bio), scrapedAt
• Live: mid, name, roomId, liveStatus, liveStatusLabel, title, online, areaName, parentAreaName, tags, liveStartedAt, coverUrl, keyframeUrl, faceUrl, liveUrl, scrapedAt

🚀 Examples

Search AI tutorials, newest first

{ "mode": "search", "searchQuery": "人工智能教程", "sortOrder": "pubdate", "maxItems": 100 }

Track a Western brand with native recall

{ "mode": "search", "searchQuery": "Nike", "searchAliases": ["AJ"], "maxItems": 200 }

Pull the danmaku off a viral video

{ "mode": "video_danmaku", "videoUrls": ["BV1rz7569EWw"], "maxDanmaku": 2000 }

Find tech creators to partner with

{ "mode": "search_users", "searchQuery": "科技", "maxItems": 50 }

Watch a gaming trend every morning

{ "mode": "search", "searchQuery": "原神", "deltaMode": true, "deltaStateKey": "genshin", "sortOrder": "pubdate" }

Check which of your KOLs are live right now

{ "mode": "live_info", "userIds": ["50329118", "672328094", "546195"] }

💻 Integrations

Feed Bilibili reactions straight into your pipeline.

Python:

from apify_client import ApifyClient

client = ApifyClient("YOUR_API_TOKEN")
run = client.actor("YOUR_USERNAME/bilibili-scraper").call(run_input={
    "mode": "video_danmaku",
    "videoUrls": ["BV1rz7569EWw"],
    "maxDanmaku": 2000,
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["text"])

JavaScript:

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_API_TOKEN' });
const run = await client.actor('YOUR_USERNAME/bilibili-scraper').call({
    mode: 'search',
    searchQuery: '人工智能教程',
    maxItems: 100,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

📈 Performance

Mode	Input	Typical result
Search	maxItems 100	Full page walk in seconds
Popular	category	Top trending in one call
Video details	BV IDs	Full metrics plus tags, fetched in parallel
Danmaku	one video	Thousands of danmaku lines, every part covered
Live status	50 creator IDs	One call

💡 Tips for best results

• Searching a Western brand? Keep auto-localize on, or add the Chinese name to searchAliases for brands outside the built-in list.
• For monitoring, set sortOrder to "Newest first" and turn on deltaMode with a distinct deltaStateKey per topic.
• For the richest reaction signal, use video_danmaku: it returns far more Gen-Z reaction text than comment counts.
• Comments are sorted best with Hot, which is the most complete public view.
• Raise maxItems for bulk pulls (category sweeps, creator monitoring, corpus building) and lower it for quick lookups.

❓ FAQ

Do I need a Bilibili account? No. It reads public data. No login, no API key, no cookies.

Can I get the actual danmaku text? Yes. The video_danmaku mode returns the scrolling-comment text itself, not just a count.

Does English search work? Yes. Auto-localize also searches the Chinese brand name and merges the results, so an English keyword still returns native videos.

Can I monitor a topic over time? Yes. Turn on delta mode and schedule the actor, and each run returns only videos it has not seen before.

What comes back for a creator? Their profile (name, followers, level, video count) plus their recent videos with full metrics.

Can I see who is live? Yes. The live_info mode checks up to 50 creators per call and returns live status, stream title, category, and current viewers for each.

Do comments include replies? Yes. Each comment thread comes back with its reply samples, linked to the parent comment by parentCommentId.

Can I get sentiment? Yes. Turn on sentiment analysis and each video, comment, and danmaku gets a polarity (positive / neutral / negative) and a score from -1.0 to +1.0.

What formats can I export? JSON, CSV, or Excel from the run's dataset.

📬 Support

Need something this scraper does not do yet? We ship features fast.

• Feature requests go straight to our backlog.
• Enterprise needs? We do custom integrations.

Response time: usually under 24 hours.