Bilibili Danmaku Scraper 哔哩哔哩弹幕提取 (No Cookie)—XML & Peak

Extract real-time XML danmaku (弹幕 / bullet comments), exact video appearance timestamps (videoTimeFormatted), high-energy peak reaction minutes (peakMinute), and sender hashes from Bilibili (哔哩哔哩 / bilibili.com) — China's premier Gen-Z video, gaming, ACG, and tech review platform.

Built for AI/LLM multimodal dialogue training corpora, video highlight detection, gaming & tech launch reception analysis, and automated RAG pipelines (n8n, Make, Dify, Claude MCP). 100% no login, no cookies, and no VPN required. Users on the Apify Free plan ($5 monthly credit) can extract up to 1,111 Bilibili danmaku bullet comments per month for free.

---

What Can This Bilibili Danmaku Scraper Do? (Features)

Bilibili Danmaku Capability	Built-In Advantage
Native XML Danmaku Stream Parsing (list.so)	Directly extracts Bilibili's real-time bullet comment pool by bvid, aid, or video URL without browser overhead or login cookies.
Peak High-Energy Minute Detection (peakMinute)	Automatically buckets bullet comments across the video timeline to pinpoint the exact 1-minute window (e.g., "02:00-02:59") where viewer reactions peak.
Smart Density Sampling (Zero Opening Spam)	Instead of only grabbing "前排" spam from the first 00:00 second, the engine samples across peak-density moments and the full video timeline sorted chronologically.
Danmaku Keyword Filtering (danmakuKeyword)	Optionally filter bullet comments by specific substrings (e.g., "高能", "泪目", "遥遥领先", "价格") to isolate targeted viewer reactions.
Automatic Keyword Video Discovery	Don't have specific BV IDs? Leave videoIds empty and enter any English or Chinese topic in searchQuery to auto-discover top videos and harvest their danmaku.
256 MB RAM Real-Time Streaming	Streams bullet records to your Apify Dataset in real-time batches of 10 rows on just 256 MB RAM, supporting scheduled runs (deltaMode) and JSON, CSV, Excel, or Parquet exports.

---

How to Scrape Bilibili Danmaku in 3 Easy Steps

1. Enter Target Videos (videoIds) — Paste one or more Bilibili BV IDs (e.g., BV1r94y1H77U), av numbers, or full video URLs; OR leave videoIds empty and enter a topic in searchQuery.
2. Set maxResults & Optional Filter — Choose how many bullet comments to extract (1 to 10,000) and optionally enter a danmakuKeyword filter.
3. Click "Start" — Download your time-aligned danmaku dataset in JSON, CSV, Excel, or Parquet, or connect via REST API, n8n, Make, or Claude MCP.

---

Bilibili Danmaku Scraper Use Cases

Buyer Profile	Bilibili Danmaku & Timestamp Intelligence Use Case
AI / LLM & Multimodal Training Teams	Build time-aligned video-to-text reaction corpora pairing exact video timestamps (videoSecond) with colloquial Chinese crowd commentary
Gaming & Consumer Tech Brands	Pinpoint which exact second of a product keynote, game trailer (Black Myth Wukong), or smartphone teardown triggered the biggest audience reaction (peakMinute)
Video Editors & MCN Agencies	Automatically detect viral highlight moments and crowd catchphrases across long-form Bilibili videos
Academic NLP & Cultural Researchers	Study real-time synchronous bullet-curtain (弹幕) linguistics, subculture slang, and viewer engagement dynamics

---

Input

All parameters are configured in the Input tab. Below is the complete input parameter reference table matching the schema and JSON input examples:

Input Field	Type	Default	Description
videoIds	array	["BV1r94y1H77U"]	Target Bilibili BV IDs, av numbers, or full video URLs. Leave empty [] to auto-discover videos via searchQuery.
danmakuKeyword	string	""	Optional substring filter to only return bullet comments containing specific words. Leave empty for full peak-density sampling.
maxResults	integer	100	Maximum number of danmaku bullet comment rows to return (1 to 10,000).
searchQuery	array	["黑神话悟空"]	Fallback keywords in English or Chinese used to auto-discover active videos when videoIds is empty.
autoLocalize	boolean	true	Automatically translates English brand/game keywords into Simplified Chinese during video auto-discovery.
deltaMode	boolean	false	Remembers scraped danmakuId values in Key-Value Store across scheduled runs so only new bullet comments are returned.
mode	string	"danmaku"	Locked to "danmaku" extraction mode.

1. Extract Peak & Timeline Danmaku from Specific Videos (JSON)

{
  "mode": "danmaku",
  "videoIds": [
    "BV1r94y1H77U",
    "https://www.bilibili.com/video/BV1xrnz6gEVD"
  ],
  "danmakuKeyword": "",
  "maxResults": 200,
  "deltaMode": false
}

2. Filter Danmaku by Keyword or Auto-Discover Videos (JSON)

{
  "mode": "danmaku",
  "videoIds": ["BV1r94y1H77U"],
  "danmakuKeyword": "高能",
  "maxResults": 100
}

---

Output

All scraped bullet comments are stored in your Apify Dataset and can be exported to JSON, CSV, Excel, XML, or Parquet. Below is a sample JSON output matching the "BV1r94y1H77U" input above:

{
  "mode": "danmaku",
  "danmakuId": "16123456789012345",
  "bvid": "BV1r94y1H77U",
  "aid": 1130123456,
  "cid": 1612345678,
  "title": "《黑神话：悟空》最终预告 | 8月20日，重走西游",
  "text": "前方高能！四大天王这压迫感绝了！",
  "videoSecond": 125.4,
  "videoTimeFormatted": "02:05",
  "peakMinute": "02:00-02:59",
  "danmakuMode": 1,
  "fontSize": 25,
  "colorHex": "#FFFFFF",
  "senderHash": "a1b2c3d4",
  "videoUrl": "https://www.bilibili.com/video/BV1r94y1H77U",
  "createdAtIso": "2024-08-08T02:18:45Z",
  "scrapedAtIso": "2026-10-28T10:30:00Z"
}

Output Field Definitions

• danmakuId: Unique Bilibili bullet comment ID (used for deduplication in deltaMode).
• bvid / aid / cid: Parent video BV ID, numeric archive ID, and video player cid.
• text: Clean bullet comment text in Simplified Chinese.
• videoSecond / videoTimeFormatted: Exact timestamp in the video where the bullet comment appears (in seconds 125.4 and MM:SS format "02:05").
• peakMinute: The 1-minute window in the parent video with the highest overall danmaku density (e.g., "02:00-02:59").
• colorHex / fontSize / danmakuMode: Visual styling metadata (#FFFFFF, font size, scrolling vs. top/bottom pinned mode).
• senderHash: Anonymized CRC32 hash of the bullet sender's account ID.

---

Automated Daily Danmaku Monitoring (Schedules & Delta Mode)

Track new bullet comments on viral videos automatically:

1. Go to Apify Console → Schedules → Create new schedule and select axon-studio/bilibili-danmaku-scraper.
2. Set deltaMode: true in your input JSON. Previously delivered danmakuId records are stored in persistent storage and skipped on future runs—so you only pay for newly posted bullet comments.

Python & Synchronous REST API Integration

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_API_TOKEN")
run = client.actor("axon-studio/bilibili-danmaku-scraper").call(run_input={
    "mode": "danmaku",
    "videoIds": ["BV1r94y1H77U"],
    "maxResults": 100,
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["videoTimeFormatted"], item["peakMinute"], item["text"])

curl -X POST "https://api.apify.com/v2/acts/axon-studio~bilibili-danmaku-scraper/run-sync-get-dataset-items?token=YOUR_APIFY_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"mode":"danmaku","videoIds":["BV1r94y1H77U"],"maxResults":50}'

---

How Much Does It Cost to Scrape Bilibili Danmaku? (Pricing)

$4.50 per 1,000 danmaku items ($0.0045 per bullet comment) + $0.02 per Actor start (Pay-Per-Event)

• Free Plan Credits: Every Apify account includes $5.00 in free monthly credits, covering up to 1,111 Bilibili danmaku bullet comments every month for free.
• Volume Discounts: Automatic lower unit prices for Apify Bronze / Starter, Silver / Scale, and Gold / Business plan subscribers.
• No Extra Server Compute Fees: Runs on just 256 MB RAM with platform compute usage fully included.
• Typical Run Costs: 100 bullet comments = $0.47 · 500 bullet comments = $2.27 · 1,000 bullet comments = $4.52.

---

FAQ & Legal Compliance

Do I need a Bilibili login or SESSDATA cookie to extract XML danmaku?

No. This Actor resolves the video's player cid and parses Bilibili's public XML danmaku stream automatically without requiring user credentials.

Is it legal to scrape public Bilibili danmaku?

This Actor extracts publicly visible bullet comments and anonymized sender hashes displayed on bilibili.com. Under GDPR and PIPL, ensure you have a lawful basis for NLP research or market analytics. Not affiliated with Bilibili Inc.

---

More Chinese Social Intelligence Scrapers by Axon Studio

• Bilibili Scraper 哔哩哔哩 (No Cookie) — All-in-One Videos, Danmaku, Comments & Creators — Full 8-mode Bilibili intelligence suite.
• Bilibili Comments & IP Geolocation Scraper — Extract Bilibili comment threads and English IP province locations (regionEn).
• Bilibili Creator Videos & UP Profile Scraper — Extract UP creator video uploads and KOL follower profiles without -352 blocks.
• Weibo Scraper 微博 (No Cookie) — Search, Posts, Comments & BlackCat — 9-mode Sina Weibo & BlackCat social listening suite.
• BlackCat Scraper 黑猫投诉 — China Consumer Complaints API — Extract Chinese consumer disputes from tousu.sina.com.cn.