Bilibili Scraper (B站) — Videos, Comments, Danmaku & Subtitles
Pricing
from $2.00 / 1,000 videos
Bilibili Scraper (B站) — Videos, Comments, Danmaku & Subtitles
Scrape Bilibili (哔哩哔哩, B站) without login: keyword search past the 1,000 cap, full video stats, UP主 profiles and full catalogs, comments, danmaku (弹幕), AI subtitles (字幕), trending and hot search. Delta mode for monitoring. B站数据采集。
Pricing
from $2.00 / 1,000 videos
Rating
5.0
(1)
Developer
Rastriq — Structured data from the world
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
8 days ago
Last modified
Categories
Features
Share
Bilibili Scraper (哔哩哔哩 / B站) — Videos, Comments, Danmaku, Subtitles & Creators
Bilibili Scraper extracts public data from Bilibili (哔哩哔哩, B站), China's largest long-form video community: keyword search, full video statistics, creator (UP主) profiles, comments with replies, danmaku (弹幕) bullet comments and every trending board — in one actor, without login or cookies.
What is Bilibili Scraper?
Bilibili is where Chinese Gen-Z discovers games, anime, tech, cars and consumer brands. With 370M+ monthly users and no YouTube in mainland China, it is the reference platform for video trends, influencer research and brand monitoring in China.
Bilibili Scraper turns that platform into clean, structured data. Instead of chaining 4–5 single-purpose actors (search, details, comments, danmaku, creators), you configure one run and get every record type in a single dataset, linked by bvid and creator_mid.
- 🔎 Search videos or creators by keyword, with sort, duration, category and date filters
- 📅 Date slicing to go past Bilibili's 1,000-results-per-search cap (tested: 1,600 unique videos for one keyword)
- 📊 Full video stats: views, likes, coins (投币), favorites, shares, danmaku and comment counts
- 👤 Creator (UP主) profiles: followers, total likes, video count, level, official verification
- 📚 Full creator catalogs: every video a creator has uploaded, newest to oldest (tested: 679 videos spanning 2013–2026 for one creator)
- 🔁 Delta mode for scheduled monitoring: only new videos, comments, danmaku and subtitles are returned and charged
- 💬 Comments and replies, with pinned comment and pseudonymised authors by default
- 🎞️ Danmaku with exact video timestamp, color, mode and send time
- 📝 Subtitles / transcripts (字幕): creator and Bilibili AI subtitles with full text and timed segments, no speech-to-text cost
- 📈 Trending: Popular (综合热门), Ranking (排行榜), Weekly must-watch (每周必看), All-time must-watch (入站必刷) and Hot search (热搜)
- 🧹 Stable schema with
source_url,scraped_at(UTC) anddata_quality_flagson every row
What data does Bilibili Scraper extract?
| Record type | Key fields |
|---|---|
🎬 video (64 fields) | bvid, aid, title, description, tags, category and parent category, duration_seconds, published_at, created_at, views, likes, coins, favorites, shares, danmaku, comments, current/highest rank, honors, pages (multi-part), width/height/orientation, paid/charging-exclusive/cooperation/interactive/360 flags, collection (合集), content_warning, recommendation_reason, staff (co-creators), short_url, creator mid/name/avatar, discovery fields |
👤 creator (36 fields) | mid, name, sign, avatar, gender, follower_count, following_count, total_likes, video_count, article_count, level, official verification (title, role, description), VIP type and label, nameplate, is_live, live_room_id, recent_videos, birthday and location when public |
💬 comment (31+ fields) | comment_id, video_bvid, parent_comment_id, parent_reply_id, dialog_id, text, like_count, reply_count, published_at, picture_urls, emotes, mentions_count, is_pinned, is_uploader_liked, is_uploader_replied, is_folded, author level/VIP/fan medal, author_hash (or full author identity when enabled) |
🎞️ danmaku | danmaku_id, video_bvid, cid, video_time_seconds, text, mode, pool, color_hex, font_size, weight, sent_at, sender_hash (every field Bilibili provides) |
📝 subtitle | video_bvid, language, language_name, is_ai_generated, ai_type, text, segments (start, end, text), covered_seconds, subtitle_url |
🔥 hot_search | rank, keyword, display_name, heat_score, heat_layer, word_type, is_commercial, trend start/end, icon_url, search_url |
🧾 Nothing is left out: besides the normalized fields, every record (except danmaku, which is fully mapped) carries a raw object with the complete original Bilibili response, so you always have every field Bilibili returns. Turn it off with includeRawData: false for lighter exports.
The dataset has ready-made views for Videos, Creators, Comments and Danmaku, so you can export each one separately to CSV, Excel or JSON.
How to use Bilibili Scraper
- Open Bilibili Scraper and click Try for free.
- Add one or more Search keywords (Chinese works best, e.g.
新能源汽车), paste video or creator URLs, or pick Trending lists. - Choose enrichments: full details, creator profiles, comments, danmaku.
- Optionally set filters: sort order, duration, category, published date range and date slicing.
- Set Max results and click Start.
- Download the results from the Output tab, or fetch them through the Apify API.
Typical use cases
- 📈 Trend and content research: which topics and formats grow in a category this week
- 🤝 KOL / influencer discovery: find UP主 by keyword and rank them by followers and engagement
- 🏷️ Brand monitoring in China: videos, comments and danmaku mentioning your brand or competitors
- 🧠 NLP and sentiment datasets: Chinese comments and danmaku with timestamps
Input parameters
| Parameter | Description | Default |
|---|---|---|
searchKeywords | Keywords to search | — |
searchType | video or user (creators) | video |
startUrls | Video URLs, BV/av IDs, creator space URLs or mids | — |
trendingLists | popular, ranking, weekly, precious, hot_search | — |
creatorVideos / maxVideosPerCreator | Full video catalog of each creator (0 = all) | false / 100 |
maxResults | Max videos, creators and hot-search rows (0 = no limit) | 100 |
maxResultsPerSource | Max per keyword or list | 100 |
fetchVideoDetails | Coins, shares, all tags, honors | true |
includeCreatorProfiles | One profile per unique uploader | false |
includeComments / maxCommentsPerVideo / commentSort | Comments, hot or newest | false / 20 / hot |
includeCommentReplies | Replies under each comment | true |
includeDanmaku / maxDanmakuPerVideo | Bullet comments | false / 1000 |
includeSubtitles / subtitleLanguages | Subtitle tracks, optional language filter (ai-zh, zh-Hans…) | false / all |
sortBy, durationFilter, categoryId | Search filters | relevance |
rankingCategory | Category for the Ranking list (0 = all) | 0 |
publishedAfter, publishedBefore, searchSliceDays | Date range and slicing | — |
deltaMode / deltaStateName | Only items not returned by previous runs of the same job | false / default |
includeRawData | Complete original Bilibili response in a raw object | true |
includeCommenterIdentity | Commenter usernames, mids, avatars, mentions and raw danmaku sender hash (pseudonymised by default, also inside raw) | false |
maxConcurrency | Parallel requests (higher = faster, more rate limiting) | 3 |
proxyUrl | Optional HTTP proxy; not needed for normal volumes | — |
Example input:
{"searchKeywords": ["新能源汽车"],"maxResults": 200,"includeCreatorProfiles": true,"includeComments": true,"maxCommentsPerVideo": 30,"publishedAfter": "2026-10-01","searchSliceDays": 7}
Output example
{"record_type": "video","bvid": "BV1iu466kEz1","url": "https://www.bilibili.com/video/BV1iu466kEz1","title": "各大新能源车企10月销量出炉,鸿蒙智行最惨淡!","tags": ["新能源汽车", "电动车", "销量"],"duration_seconds": 2410,"published_at": "2026-11-06T12:58:19Z","creator_mid": 3461565976152389,"creator_name": "绝丗篮球","view_count": 39380,"like_count": 381,"coin_count": 18,"favorite_count": 173,"share_count": 32,"danmaku_count": 204,"comment_count": 185,"discovery_source": "search","discovery_keyword": "新能源汽车","discovery_rank": 1,"detail_level": "full","scraped_at": "2026-11-07T21:26:23Z","data_quality_flags": []}
{"record_type": "danmaku", "video_bvid": "BV1iu466kEz1", "video_time_seconds": 3.957, "text": "鸿蒙智行是哪家牌子呀", "mode": "scroll", "color_hex": "#ffffff", "sent_at": "2026-11-07T14:53:06Z", "sender_hash": "ffb4c17b"}
{"record_type": "comment", "video_bvid": "BV1iu466kEz1", "comment_id": "323175400336", "parent_comment_id": null, "text": "骚操作一大堆……", "like_count": 110, "reply_count": 22, "published_at": "2026-11-07T00:10:34Z", "author_hash": "bb7badaf5cf8bf82", "author_level": 6, "is_pinned": false}
{"record_type": "creator", "mid": 546195, "name": "老番茄", "follower_count": 20875907, "video_count": 684, "level": 6, "official_verified": true, "url": "https://space.bilibili.com/546195"}
{"record_type": "subtitle", "video_bvid": "BV1Rz4n6nEjp", "language": "ai-zh", "is_ai_generated": true, "segment_count": 598, "covered_seconds": 838.52, "text": "女人刑满释放……", "segments": [{"start": 0.5, "end": 2.1, "text": "女人刑满释放20年的牢狱"}]}
Data quality flags
Every record has data_quality_flags. Records are never dropped: flags tell you what to check, and you decide what to filter.
| Flag | Meaning |
|---|---|
views_lagging | More likes than views: Bilibili updates view counts with a delay on very recent videos |
view_count_rounded | View count comes rounded (e.g. 143.9万) because full details were off; enable fetchVideoDetails for exact numbers |
missing_title, missing_view_count, missing_published_at, missing_creator | The source did not return that field |
empty_subtitle / empty_text | Subtitle track or comment with no text |
detail_level tells you whether a video has full statistics (full) or listing-level data (listing). discovery_source, discovery_keyword and discovery_rank tell you where each video was found. A RUN_SUMMARY record in the run's key-value store lists record counts, requests, retries and any errors by type.
Good to know
- Listing-level videos (search without full details, creator catalogs) do not include coins, shares,
cidor tags; turn onfetchVideoDetailsto get them. publish_location(IP属地) is only provided by Bilibili on trending lists.- Danmaku and subtitles are collected for the first part (P1) of multi-part videos. Danmaku come from Bilibili's current danmaku pool for the video, which Bilibili caps (usually the latest few thousand).
- Subtitles and full creator catalogs are occasionally unavailable; when that happens the run continues, subtitles are skipped and catalogs fall back to the creator's recent videos, and the
RUN_SUMMARYnotes it. - Not every video has subtitles; Bilibili AI subtitles (
ai-zh) are the most common.
💰 How much does it cost to scrape Bilibili?
Bilibili Scraper uses pay-per-event pricing: you pay per record returned (video, creator, comment, danmaku, subtitle track, hot search), never for empty results. Current prices are shown in the Pricing tab.
How do I keep costs under control? Start with maxResults: 20, then scale up. Use the run's maximum-charge setting to cap spend: Bilibili Scraper stops cleanly when the limit is reached. With delta mode, scheduled runs only pay for new items.
Related Actors
- YouTube Transcript Scraper: the YouTube side of your video research
- TikTok Trending Videos Scraper: trending short video outside China
FAQ
Do I need a Bilibili account or cookies? No. You don't need an account, cookies or a proxy to use Bilibili Scraper.
How many comments can I get per video?
Bilibili Scraper pages through top-level comments (hot or newest order) and their replies, up to maxCommentsPerVideo. Very deep pagination on huge threads may end earlier than the full comment_count, which is always exact.
Do all videos have subtitles?
No. Many Bilibili videos have AI-generated Chinese subtitles (ai-zh) and some have creator subtitles; videos without any return no subtitle record.
How do I get more than 1,000 results for one keyword?
Set publishedAfter and searchSliceDays. Bilibili Scraper runs one search per date window and de-duplicates results.
How does delta mode work?
Bilibili Scraper remembers the IDs it already returned in a key-value store in your account (bilibili-delta-<name>). The first run is the baseline; later runs output and charge only new items. For comment monitoring, use commentSort: newest. Use a different deltaStateName per monitoring job.
Are commenters' identities included?
It's your choice. By default authors and danmaku senders are replaced by stable pseudonymous hashes (author_hash, sender_hash). Enable includeCommenterIdentity only if you have a lawful basis to process personal data.
Is scraping Bilibili legal? Bilibili Scraper collects publicly available data only. You are responsible for complying with Bilibili's terms of use, robots.txt, and data protection laws such as China's PIPL and the EU GDPR.
Something is missing or broken? Open an issue in the Issues tab and we will look into it.
🇨🇳 中文说明
Bilibili Scraper 是什么?
Bilibili Scraper 是一个哔哩哔哩(B站)数据采集工具,无需登录、无需 Cookie、无需代理。一次运行即可获取视频搜索结果、完整视频数据、UP主资料与全部投稿、评论及回复、弹幕、字幕(AI 字幕与 UP 主上传字幕)以及热门榜单和热搜。
可以采集哪些数据?
| 类型 | 主要字段 |
|---|---|
视频 video | BV号、标题、简介、标签、分区、时长、发布时间、播放、点赞、投币、收藏、分享、弹幕数、评论数 |
UP主 creator | mid、昵称、粉丝数、获赞数、投稿数、等级、认证信息 |
评论 comment | 评论内容、点赞数、回复数、发布时间、是否置顶(默认匿名化) |
弹幕 danmaku | 弹幕内容、出现时间点、颜色、模式、发送时间 |
字幕 subtitle | 语言、是否 AI 生成、全文、带时间轴的分段 |
热搜 hot_search | 排名、关键词、热度 |
如何使用?
- 输入搜索关键词(如
新能源汽车)、视频链接 / BV号,或 UP主空间链接 / mid。 - 选择需要的数据:完整详情、UP主资料、全部投稿、评论、弹幕、字幕、热门榜单。
- 可按排序、时长、分区和发布时间筛选;按日期分段可突破单个关键词 1000 条的限制。
- 开启「仅新增(delta)」模式,定时运行时只返回并计费新增数据。
- 点击 Start,导出 JSON、CSV 或 Excel,或通过 Apify API 获取。
如何收费?
按结果计费(pay-per-event):只为返回的记录付费,具体价格见 Pricing 页面。可设置最高费用上限,达到上限后自动停止。