Bilibili Video Danmaku Extractor

The Problem

Bilibili danmaku is not only real-time commentary, but also context, jokes, and viewer reactions. The web player is convenient for watching, but it is not ideal for preserving, searching, or analyzing all danmaku. If you want to save a video’s danmaku locally for content review, case notes, or structured statistics, manual copying often loses fields such as time, type, and color, and it is hard to keep a clean chronological order.

How It Works

This skill wraps Bilibili’s public endpoints in a lightweight extraction flow:

• accepts a BV ID or video URL
• reads video metadata, including title, uploader, duration, and cid
• requests the danmaku list with cid, covering rolling, bottom, and top types
• handles gzip-compressed responses and exports JSON plus time-ordered Markdown

The JSON output keeps full video metadata and danmaku records; the Markdown output is easier to read and preview. Fields include text, time, type, type_code, color, and timestamp, which helps distinguish rolling, bottom, top, reverse, and other danmaku types.

Boundaries

• It targets public videos; the material says no login is required, but Bilibili risk controls and request-rate limits still apply
• The number of extracted danmaku depends on the video; private content and cross-platform data are not covered
• If the result is empty, or you see 412 / 403, check link validity, network state, and slow down retries