YouTube Bilibili AI Summarizer（视频字幕摘要器）

About

Video Caption AI Summarizer turns accessible video captions and transcripts into clear, reusable AI summaries. It is built for people who learn, research, review, and work from long-form video on common video and course pages.

Instead of rewatching a long lecture, interview, product demo, technical talk, or course video just to find the key points, you can read the available captions, send them to the AI service you configure, and copy structured Markdown notes from the video page.

Main features:

- Summarize captions from supported video pages and generic HTML5 video pages.
- Read available caption tracks, VTT/SRT subtitles, visible transcript text, or manually pasted transcript content.
- Bring your own AI API or local model endpoint instead of being locked into one provider.
- Save multiple API profiles with provider type, endpoint, API key, model name, temperature, and output token limits.
- Start quickly with preset configuration templates in Settings, then edit the endpoint and model details for your own account.
- Customize summary prompts, output templates, summary language, tone, structure, and detail level.
- Split long transcripts into chunks, summarize each part, then merge the final result.
- Copy generated Markdown summaries or original transcript text for notes, research, documentation, and knowledge-base workflows.
- Control privacy details such as timestamp sharing, sensitive term redaction, and optional local summary history.
- Use light, dark, or automatic theme settings for the floating video-page panel.

Good for:

- Students and self-learners summarizing lectures, tutorials, course videos, and conference talks.
- Researchers and analysts reviewing interviews, presentations, webinars, and recorded discussions.
- Creators and product teams turning video material into outlines, notes, briefs, or reusable references.
- Developers and AI users who prefer using their own API key, local service, or compatible endpoint.

Privacy and data handling:

The extension sends the selected caption or transcript text to the AI service configured by the user. API keys and settings are stored in local browser extension storage. If local history is enabled, recent summaries are stored locally in the browser. Users can disable timestamp sharing and configure sensitive terms to redact before sending text to their chosen AI service.

Important limitations:

The extension can only summarize captions or transcript text that the browser can access. It may not read captions that are burned into the video image, protected by DRM, hidden behind unavailable account permissions, or changed by a platform update. When automatic caption detection is unavailable, users can paste transcript text manually.

---

中文介绍：

Video Caption AI Summarizer 可以把视频字幕和转写文稿整理成清晰、可复用的 AI 摘要，适合经常通过视频页面、课程平台和技术分享学习、研究或工作的用户。

你不必反复拖动长视频查找重点。插件会读取浏览器可访问的字幕轨道、VTT/SRT 字幕、页面可见的 transcript 文本，或你手动粘贴的转写内容，并调用你自己配置的 AI API 生成结构化 Markdown 摘要。

核心能力：

- 支持常见视频页面和通用 HTML5 视频页面。
- 支持字幕轨道、VTT/SRT 字幕、页面可见转写文稿和手动粘贴文本。
- 使用你自己的 AI API 或本地模型端点，不绑定单一服务商。
- 可保存多个 API 配置，在不同端点、密钥、模型和参数之间切换。
- 内置配置模板，可快速填入常见端点和模型字段，并按你的账号继续调整。
- 支持自定义总结 Prompt、输出模板、摘要语言、结构和细节密度。
- 长字幕会先分段提炼，再合并成最终摘要。
- 可复制 Markdown 摘要或原始字幕，用于学习笔记、研究记录、文档和知识库。
- 支持时间戳发送开关、敏感词替换和可选的本地历史记录。
- 支持浅色、深色和自动主题的视频页浮动面板。

隐私说明：

插件会把用户选择的字幕或转写文本发送到用户自己配置的 AI 服务。API Key 和设置保存在浏览器扩展本地存储中。如果启用历史记录，最近摘要会保存在本机浏览器中。用户可以关闭时间戳发送，并配置敏感词在发送前自动替换。

使用边界：

插件只能处理浏览器能够访问到的字幕或文本。对于烧录在视频画面中、受 DRM 或账号权限限制、或平台结构变化导致无法读取的字幕，可能无法自动总结。遇到这种情况，可以使用手动粘贴转写文本的方式。