Bilibili Scraper (Videos, Comments, Uploaders)

Video data from Bilibili by keyword, video link, popular list or ranking: title, description, tags, category, duration, publish date, uploader and every public stat (views, likes, coins, favourites, shares, danmaku, comments). Add top comments with their replies, and uploader profiles with fan, video and like counts. Pay per video, comment and uploader delivered. No login, no browser, no API key.

This one speaks the site's own signed API (the same calls the web app makes), rotates its identity before Bilibili's risk control bites, retries comments on a residential IP when Bilibili hands a datacenter an empty list (with the fallback on), and ends every run with a status line that says what came back and why anything is missing.

What you get

Video rows (type: video): bvid, aid, title, description, url, cover, duration_seconds, published_at, category, tags (keyword searches only; Bilibili's detail and list answers carry no tags), uploader_mid, uploader_name, uploader_url, views, likes, coins, favourites, shares, danmaku, comments (the full count), rank (for lists), source.

Comment rows (type: comment): rpid, bvid, aid, text, author, author_mid, likes, reply_count, created_at, nested replies.

Uploader rows (type: uploader): mid, name, url, face, sign, fans, following, videos, likes, level, official_title.

Use cases

• Trend research: what is popular on Chinese video right now, by category.
• Keyword monitoring: the videos Bilibili ranks for a product, game or topic (relevance order), with engagement counts.
• Creator analysis: an uploader's profile and public stats; their videos via a name search.
• Audience research: top comments and replies, with likes, for sentiment work.
• Datasets: titles, tags and stats for recommendation or NLP projects.

How to use it

1. Type keywords, paste video links or BV ids, add uploader ids, or turn on the popular or ranking list.
2. Choose how many videos per source and whether to include comments.
3. Start the run. Download as JSON, CSV or Excel, or read the dataset through the API.

What each source gives

Source	Per request	How far	Notes
Keyword search	20 videos	up to 1,000 (50 pages)	Bilibili matches loosely and never answers "0 results"; the status line shows how many came back
Video link or id	the full record		a removed or restricted video costs nothing and the status line says why
Popular list	20 a page	up to 400 (20 pages)
Ranking	100		by category
Uploader id	the profile		the uploader's own video list needs a browser-grade signature Bilibili does not hand out; search the name instead
Comments	the 3 top comments, then up to 20 replies under each	about 63 per video at most	Bilibili shows that much to visitors without a login; the full count is on the video row. Retried on residential (fallback on) when a datacenter IP gets an empty list

Cost

Delivered	Cost
100 videos	about $1.00
1,000 videos	about $10
1,000 comments	about $2.00
100 uploaders	about $1.00

A small fee per run plus a fee per video, per comment and per uploader (the table assumes $0.01, $0.002 and $0.01). A video found by two of your sources is charged once, and so are its comments. Removed or restricted videos, refused requests and empty searches are free.

Scheduling and API

curl -X POST "https://api.apify.com/v2/acts/oski~bilibili-scraper/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"keywords": ["原神"], "maxVideos": 100, "includeComments": true, "maxCommentsPerVideo": 20}'

If something looks wrong

Status line says	What it means	What to do
Input problem: Add at least one Bilibili video link ...	Nothing usable in the inputs	Add a keyword, a BV id or an uploader id
Skipped '...': not a Bilibili video link	A line was not Bilibili	Remove it
The video does not exist or was removed	Bilibili answered -404	Nothing to do; the row is free
Bilibili restricts this video	Region, age or membership gate	Nothing to do; the row is free
Bilibili refused every identity and proxy we tried	Risk control on both pools	Run again later
Uploader video lists are not available yet	Bilibili's space listing needs a browser signature	Search the uploader's name as a keyword
"python" stopped at your maximum of 200 videos (1,000 listed)	Your cap	Raise Videos per source
Bilibili shows visitors without a login only the 3 top comments ...	Bilibili's visitor view	Nothing to do; the comments column on the video row has the full count
"python": Bilibili refused the later pages; 340 videos were kept	Risk control part way	Run again later for the rest

Frequently asked questions

Is the text in Chinese? Titles, descriptions, tags and comments are whatever the uploader wrote, mostly Chinese. Category names are Bilibili's Chinese names.

Can I get a whole channel? Not yet: the uploader video list endpoint refuses non-browser clients. A keyword search for the uploader's name gets most of their recent videos.

Why only a few comments per video? Bilibili shows visitors without a login the 3 top comments and the first 20 replies under each. Logging in is not something this actor does.

Does it download videos? No. It reads public metadata, stats and comments only.

Related actors by the same developer

• YouTube Transcript Scraper
• TikTok Transcript Scraper
• Telegram Channel Scraper