Bilibili Video & Comment Scraper (CN)
Under maintenancePricing
$2.00 / 1,000 result items
Bilibili Video & Comment Scraper (CN)
Under maintenanceLogin-less Bilibili (bilibili.com) video and comment scraper: video search, video details, top-level comments, user profiles and uploader videos. No account needed; guest-access limits are documented.
Pricing
$2.00 / 1,000 result items
Rating
0.0
(0)
Developer
MYM
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
2 days ago
Last modified
Categories
Share
Bilibili Scraper (CN)
Login-less Bilibili (bilibili.com) data via a private scraping backend — this Actor is a thin bridge and contains no scraping code.
Modes
| Mode | Input | Output |
|---|---|---|
search | keyword, pages (≤10), order, duration | video rows: title, url, duration, views, likes, coins, favorites, replies, danmaku, tags, uploader |
video_detail | videos (URL / BV id / aid) | per-video detail: tags, reply count; full view stats when the network allows |
video_comments | videos, maxComments (≤500) | top-level comments with author, likes, location, first replies |
user_profile | users (space URL / mid) | name, avatar, sign, level, followers, following, likes received, video count |
user_videos | users, optional name, pages, maxVideos | that uploader's videos found via search, exact-id filtered |
Known limitations (measured, 2026-09-20)
- Comments without login: from datacenter networks Bilibili returns only the top ~3 comments per video even when the real count is huge. Full comment paging needs login cookies or clean/residential egress — not included.
- Video detail is partial on datacenter egress: the view-statistics endpoint is risk-control blocked (HTTP 412 / guest-masked -404). Items still include tags and the exact reply count; the blocked part is reported in run errors.
- user_videos is NOT an exhaustive listing: it searches the uploader's name and keeps only exact uploader-id matches. The direct uploads listing endpoint is blocked on datacenter egress.
- Rate limiting: rapid search bursts intermittently get HTTP 412. Requests are paced (~0.35–0.6 s gap) and retried with bounded backoff (max 3 attempts/request), then fail with a classified error.
- No login, no cookies, no danmaku download, no media files.
Input
See the Input tab — the schema mirrors the backend contract 1:1. Invalid input (wrong mode, missing keyword/lists) fails the run immediately, before any scraping, so you never pay for a malformed request.
Output
One dataset item per video / comment / user. Run notes (per-input errors and the limits above) are printed to the run log and stored in the STATS value, not billed as items.
