Bilibili Video & Search Scraper

Bilibili Scraper

Scrape public video, creator, relation, comment, danmaku, subtitle, and keyword-search data from bilibili.com with one operation per run. Free plan pricing: $0.002 per dataset row plus $0.00005 per apify-actor-start event.

• Search public video cards by keyword with titles, URLs, creators, counts, durations, and publication dates.
• Collect creator profiles, relation statistics, and creator videos for creator research.
• Extract comments and danmaku for audience-feedback analysis.
• Collect video details and subtitle metadata for cataloging.
• Run without signing in or an API key.

What you get

Each run uses one operation. Rows contain source-backed values and normalized aliases; unavailable fields are omitted.

Field family	Fields	Type	What it contains
Run metadata	_operation, _fetchedAt, status	string or ISO 8601 string	Operation, UTC time, and row status
Pagination and provenance	_page, _sourceKeyword, _sourceAid, _sourceCid	integer or string	Page number and source IDs
Video identity	title, url, videoTitle, videoDesc, bvid, aid, cid	string	Titles, descriptions, IDs, and canonical URL
Video page assets	videoPageUrl, coverUrl	string	Video and cover URLs
Timing and publication	durationSec, durationStr, publishedAt, publishedDate, createdAt	integer or string	Duration and publication or creation values
Category and tags	videoTypeId, videoTypeName, tags	integer, string, or array of strings	Categories and page tags
Video metrics	playCount, playCountDisplay, likeCount, coinCount, favoriteCount, shareCount, commentCount, danmuCount, danmuCountDisplay	integer or string	Video counts and display labels
Video parts	videoPageCount, videoPages	integer or array	Listed parts and part data
Creator profile	userId, userName, userPageUrl, avatarUrl, bio, sex, school, profession, level, userVipStatus, verification	string or integer	Creator or commenter identity and profile fields
Creator relations	followerCount, followingCount, whisperCount, archiveCount	integer	Relation and archive statistics
Comment identity	commentId, parentCommentId, rootCommentId, userLevel	string or integer	Comment and author identifiers
Comment content	commentText, replyCount, commentTime	string or integer	Text, replies, and creation time
Danmaku data	danmuId, danmuText, timeOffsetSec, mode, fontSize, color, pool, userHash	string, number, or integer	Text, timing, display values, and source hash
Subtitle metadata	subtitleId, subtitleLanguage, subtitleName, subtitleType, subtitleUrl	string or integer	Subtitle IDs, languages, type, and URL

Fields by operation

Every operation returns the run metadata fields plus the families below, when the source provides them.

Operation	Field families returned
searchVideo	Pagination and provenance, video identity and assets (title, url, bvid, coverUrl), creator userId, userName, userPageUrl, play and danmaku counts with display labels, duration, publishedDate
videoDetail	Video identity, aid, cid, description, cover, duration, publishedAt, createdAt, category, the seven video counts, video parts, tags, creator profile and relation fields from the page
userDetail	Creator profile fields, publishedAt, archiveCount, followerCount, followingCount, whisperCount
userRelationStat	userId, url, userPageUrl, followerCount, followingCount, whisperCount
userVideos	Video identity, aid, description, cover, duration, publishedAt, playCount, commentCount, videoTypeName, creator identity
videoComments	Comment identity and content, likeCount, replyCount, commentTime, commenter userId, userName, userPageUrl, userLevel, userVipStatus, video aid and URLs
videoDanmu	Danmaku data, createdAt, aid, cid, _sourceAid, _sourceCid, URLs
videoCaption	Video identity (bvid, aid, cid, videoTitle), URLs, subtitle metadata

publishedAt, createdAt, and commentTime are Unix seconds. searchVideo playCount and danmuCount are approximate: the Actor derives them from the rounded labels on search cards (for example 2079.9万), and the exact label is kept in playCountDisplay and danmuCountDisplay. In videoDetail the counts come from the video page's stat block.

Sample output

This shortened sample omits duplicate video-title and URL fields, plus cover and profile URL fields. All shown values are copied from one real dataset record.

{
  "_operation": "searchVideo",
  "_fetchedAt": "2026-10-15T01:25:24.113157+00:00",
  "status": "success",
  "_page": 1,
  "title": "【全748集】目前B站最全最细的Python零基础全套教程，2026最新版，包含所有干货！七天就能从小白到大神！少走99%的弯路！存下吧！很难找全的！",
  "url": "https://www.bilibili.com/video/BV1rpWjevEip/",
  "bvid": "BV1rpWjevEip",
  "_sourceKeyword": "python",
  "userId": "3546597933714079",
  "userName": "Python官方课程",
  "playCountDisplay": "2079.9万",
  "playCount": 20799000,
  "danmuCountDisplay": "13.3万",
  "danmuCount": 133000,
  "durationStr": "39:58:14",
  "durationSec": 143894,
  "publishedDate": "2024年8月22日"
}

Pricing

Billing uses pay-per-event pricing. A dataset row is billed once as one scraped result with all fields.

Charge event	Billing unit	Price per event (Free plan)	Condition
apify-actor-start	One Actor Start event	$0.00005 (Free plan)	Charged when the Actor starts running
apify-default-dataset-item	One dataset row	$0.002 (Free plan)	One scraped result with all fields

On the Free plan, 1,000 row events cost $2, and 10,000 cost $20; 1,000 start events cost $0.05.

2,499 x $0.002 = $4.998 on the Free plan, so the $5 monthly Free credit covers about 2,499 rows. At the default maxResults of 50, the Free credit covers 49 full runs (2,450 dataset rows) with one apify-actor-start event each; additional start events reduce that count.

Only the events listed above are charged. This Actor does not use the Pay per event + usage option, so compute, proxy traffic, and storage are not billed separately.

How to use

1. Choose one operation: videoDetail, userDetail, userRelationStat, userVideos, videoComments, videoDanmu, videoCaption, or searchVideo.
2. Enter the matching identifier: keyword for searchVideo; bvid for videoDetail or videoCaption; userId for creator operations; aid for comments; and both aid and cid for danmaku.
3. Set maxResults for the row cap and maxPages for search, creator-video, or comment pagination.
4. Select an optional Apify proxy configuration.
5. Start the run and inspect the resulting dataset rows.
6. Repeat the run through Apify Schedules; daily is a starting point for tracking keyword search results or creator videos.

{
  "operation": "searchVideo",
  "keyword": "python",
  "maxResults": 10,
  "maxPages": 2,
  "proxyConfiguration": {
    "useApifyProxy": false
  }
}

Export the dataset as JSON or CSV, or read it through the Apify dataset API.

Input parameters

Parameter	What it does	Default
operation	Selects the collection mode: videoDetail, userDetail, userRelationStat, userVideos, videoComments, videoDanmu, videoCaption, or searchVideo.	searchVideo
keyword	Search term for searchVideo; Chinese, English, and mixed terms are accepted.	Not set
bvid	Bilibili BV identifier for videoDetail and videoCaption; expects BV plus 10 letters or digits.	Not set
aid	Positive numeric video ID for videoComments and videoDanmu.	Not set
cid	Positive numeric video-part ID for videoDanmu.	Not set
userId	Positive numeric creator ID for userDetail, userRelationStat, and userVideos.	Not set
maxResults	Maximum dataset rows; the schema accepts 1 through 1000.	50
maxPages	Page limit for search, creator videos, and comments; the schema accepts 1 through 50.	5
proxyConfiguration	Optional proxy settings. The input form prefills the Apify proxy with the RESIDENTIAL group ({"useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"]}). If you omit the field or set useApifyProxy to false, requests go out over a direct connection, as in the example input above.	Prefilled: Apify proxy, RESIDENTIAL

Limits and edge cases

Records are built before output writing. After collection, rows are pushed one at a time; a failure before writing leaves no rows, while an abort or timeout during writing leaves rows already written.

Condition	What the buyer sees	What to do
maxResults is reached	No more than the selected row cap is written.	Increase maxResults within the schema range.
A paginated operation reaches maxPages	Search, creator-video, or comment collection stops at the configured limit.	Increase maxPages up to 50 and adjust maxResults.
Pagination reaches the source end or returns no page rows	Fewer rows than requested, or no rows, may be written.	Check the keyword or identifier and try again.
Bilibili returns a challenge page or HTTP 403, 406, or 412	The request is retried up to three attempts and is never treated as data. If it still fails, the run fails; userDetail can fall back to the public space page when the profile or relation API errors.	Try an optional Apify proxy configuration or rerun later.
A video detail page has no usable videoData, a creator response has no creator name, or relation statistics have no usable data	The affected operation fails.	Verify the Bilibili identifier and rerun.
Danmaku returns a non-XML response or unparsable XML	The danmaku operation fails.	Verify aid and cid, then rerun.

The schema accepts maxResults from 1 through 1000. The execution path also returns zero rows when it receives maxResults equal to 0, although that value is outside the schema range.

FAQ

How much does it cost to scrape Bilibili?

The Actor charges $0.002 per dataset row on the Free plan; see Pricing for the full calculation.

Do I need a Bilibili account or API key?

No. The Actor does not sign in and does not require a Bilibili account or API key. You only need an Apify account to run it.

Why did my run return no rows?

A keyword may have no matching rows, or creator-video and comment collection may reach the source end. Danmaku and subtitle metadata operations may also return no rows when the public source provides no corresponding records. Check the operation, identifier, and pagination settings before rerunning.

What happens if Bilibili returns an anti-bot challenge?

The Actor detects challenge pages, retries, and does not write the page as a collected record; if retries fail, the run fails. Try an optional Apify proxy configuration or run the collection again later.

Is it legal to scrape Bilibili public data?

The Actor reads public Bilibili page data and public endpoints. Whether a particular use is allowed depends on Bilibili’s terms of use and applicable law for your use case; the Actor does not provide a legal guarantee.

Related actors

• Douyin Scraper — related Chinese social-video data collection for hot trends and videos.
• Bluesky Scraper — collects Bluesky posts, profiles, and followers.
• Facebook Comments Scraper — related public comment collection for Facebook.