Bilibili Comments
Fetch Bilibili video comments in bulk, including nested replies. Requires a Bilibili cookie for authentication.
Bilibili-Comments-MCP
一个基于 Model Context Protocol (MCP) 的 B 站视频评论获取工具.
快速开始
1. clone本项目
git clone https://github.com/222wcnm/Bilibili-Comments-MCP.git
cd Bilibili-Comments-MCP
2. 安装依赖
npm install @modelcontextprotocol/sdk axios
3. 配置客户端(如Claude客户端)
在 MCP 客户端的配置文件中添加:
{
"mcpServers": {
"bilibili-comments": {
"command": "node",
"args": ["/path/to/bilibili_mcp.js"],
"env": {
"BILIBILI_SESSDATA": "your_bilibili_sessdata_here"
}
}
}
}
环境变量
配置方式
-
BILIBILI_SESSDATA
:Bilibili Cookie 中的 SESSDATA 值。- 获取方式:登录 Bilibili 网站,打开浏览器开发者工具 (F12),在 Network (网络) 选项卡中刷新页面,找到任意一个
bilibili.com
的请求,在 Request Headers 中找到 Cookie,提取SESSDATA=xxx
部分的值。
- 获取方式:登录 Bilibili 网站,打开浏览器开发者工具 (F12),在 Network (网络) 选项卡中刷新页面,找到任意一个
工具功能
get_video_comments
获取 B 站视频评论,支持分页、排序和楼中楼回复。
参数:
bvid
/aid
- 视频ID(二选一)page
- 页码,默认1pageSize
- 每页数量(1-20),默认20sort
- 排序:0按时间,1按热度includeReplies
- 是否包含楼中楼回复,默认trueoutputFormat
- 输出格式:markdown 或 json,默认markdowncookie
- B站Cookie(可选)
示例(Markdown格式):
{
"bvid": "BV1xx411c7mD",
"page": 1,
"pageSize": 20,
"sort": 1,
"includeReplies": true,
"outputFormat": "markdown"
}
示例(JSON格式):
{
"bvid": "BV1xx411c7mD",
"page": 1,
"pageSize": 20,
"sort": 0,
"includeReplies": false,
"outputFormat": "json"
}
get_dynamic_comments
获取 B 站动态评论,支持分页和楼中楼回复。
参数:
dynamic_id
- 动态ID(必需)page
- 页码,默认1pageSize
- 每页数量(1-20),默认20includeReplies
- 是否包含楼中楼回复,默认trueoutputFormat
- 输出格式:markdown 或 json,默认markdowncookie
- B站Cookie(可选)
示例:
{
"dynamic_id": "123456789",
"page": 1,
"pageSize": 10,
"includeReplies": true,
"outputFormat": "markdown"
}
Cookie 获取
- 登录 B 站网页版
- 打开开发者工具 (F12)
- 切换到 Network 标签
- 刷新页面,找到任意请求
- 复制 Request Headers 中的 Cookie 值
动态ID获取方法
- 在B站手机App中打开想要获取评论的动态
- 点击分享按钮
- 选择"复制链接"
- 链接格式通常为:
https://t.bilibili.com/动态ID
- 提取链接中的数字部分作为dynamic_id参数
常见问题解决
动态评论获取失败
- 错误代码-404: 动态不存在或已被删除,请检查dynamic_id是否正确
- 错误代码-101: Cookie已过期,请重新获取并更新SESSDATA
- 错误代码-403: 访问权限不足,某些动态可能需要登录才能查看评论
Related Servers
MCP Webscan Server
Fetch, analyze, and extract information from web pages.
ScreenshotOne
Render website screenshots with ScreenshotOne
Query Table
A financial web table crawler using Playwright that queries data from multiple websites with fallback switching.
MCP Undetected Chromedriver
Automate Chrome browser control while bypassing anti-bot detection using undetected-chromedriver.
Playwright MCP
Browser automation using Playwright, enabling LLMs to interact with web pages through structured accessibility snapshots.
MCP YouTube Transcript Server
Retrieves transcripts from YouTube videos for content analysis and processing.
Headline Vibes Analysis
Analyzes the sentiment of news headlines from major US publications using the NewsAPI.
YouTube Transcript
A zero-setup server to extract transcripts from YouTube videos on any platform.
LinkedIn Profile Scraper
Fetches LinkedIn profile information using the Fresh LinkedIn Profile Data API.
Selenium MCP Server
Control web browsers using the Selenium WebDriver for automation and testing.