Frequently Asked Questions

Yes. Xiaoliu BOT danmuku is currently a free public tool. Danmaku download, charts, and AI analysis can be used without a paid plan. During busy periods, built-in AI analysis may queue, time out, or fail because it depends on server load and upstream model availability.

The site supports public Bilibili videos that can be accessed normally. You can paste a BV ID, a standard video URL, or a URL with query parameters. Common inputs include BV1w7411Y77t and https://www.bilibili.com/video/BV1w7411Y77t.

Yes. Paste the Bilibili video page URL into the search box and the site will extract the BV ID automatically. A full video page URL usually gives the best success rate.

After parsing, you can download CSV and TXT files. CSV keeps richer danmaku fields for spreadsheet filtering, statistics, and further analysis. TXT is lighter and easier to read, copy, or pass into text workflows.

The files include danmaku text and timing information. TXT is simplified for reading and AI analysis, while CSV is better for complete field-level statistics and custom filtering.

The density chart helps locate high-interaction moments in a video. The word cloud summarizes recurring audience keywords. Together they are useful for video review, topic research, and audience-response analysis.

The available data depends on what can be publicly fetched. New videos, low-danmaku videos, access restrictions, or temporary upstream issues may reduce the data shown here.

Danmaku analysis uses comment data to summarize audience attitude, interaction patterns, and hot topics. Subtitle deep analysis combines an uploaded TXT subtitle with danmaku data, which is better for reviewing structure, key moments, and content strategy.

Yes. Danmaku content analysis works without subtitles. Subtitle deep analysis requires a TXT subtitle file because it reads both the subtitle text and the danmaku data.

No. Your custom Base URL, model name, and API key are stored only in your browser's local storage. When custom API mode is enabled, the browser calls your OpenAI-compatible endpoint directly. Clearing browser data, switching devices, or using private browsing can remove the configuration.

Built-in AI uses the server-side model configuration and requires no key from you, but tasks share the public queue. With custom API mode, the server prepares danmaku, subtitle text, and prompts, while your browser calls the endpoint you configured. This can improve model choice, speed, queueing, context length, and long-video capacity if your endpoint supports it.

A shared report saves a structured snapshot of the current result page so video info, charts, and AI analysis can be reopened later. Raw danmaku files and uploaded subtitle text are not kept as long-term public report data.

Shared reports have a retention period, currently about 30 days depending on site configuration. For long-term sharing, keep the standard result URL as well, such as /result?bvid=BV_ID.