Extracts subtitles, searches keywords with timestamps, retrieves video metadata, and lists available subtitle languages from Bilibili videos.
Extracts subtitles (manual and auto-generated), searches keywords with timestamps, retrieves video metadata (title, duration, description), and lists available subtitle languages from YouTube videos.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@subtitle-analyzer extract subtitles from https://www.youtube.com/watch?v=dQw4w9WgXcQ"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
基于 yt-dlp 的字幕分析 MCP 服务器，支持 YouTube 和 Bilibili 平台。
🎬 字幕提取 - 从视频提取字幕（支持手动字幕和自动生成字幕）
🔍 时间戳搜索 - 搜索关键词并定位到具体时间点
📋 视频信息 - 获取视频标题、时长、描述等元信息
🌍 多语言支持 - 支持中文、英文、日文等多种语言字幕
Related MCP server: youtube-mcp
Python 3.10+
yt-dlp（系统级安装）
# 安装 yt-dlp
pip install yt-dlp
# 或者使用 brew (macOS)
brew install yt-dlp# 克隆或下载项目
cd subtitle-analyzer-mcp
# 安装依赖
pip install -e .编辑配置文件：
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"subtitle-analyzer": {
"command": "python",
"args": ["-m", "src.server"],
"cwd": "/path/to/subtitle-analyzer-mcp"
}
}
}或者使用 uv（推荐）：
{
"mcpServers": {
"subtitle-analyzer": {
"command": "uv",
"args": ["run", "--directory", "/path/to/subtitle-analyzer-mcp", "python", "-m", "src.server"]
}
}
}# 添加 MCP 服务器
claude mcp add subtitle-analyzer -- python -m src.server --cwd /path/to/subtitle-analyzer-mcpYouTube 和 Bilibili 的部分视频需要登录才能访问。由于 MCP 服务器运行在后台进程中，无法直接访问浏览器 cookies，因此需要预先导出 cookies 文件。
系统按以下优先级查找认证信息：
cookies_file 参数 - 工具调用时直接指定
YT_DLP_COOKIES 环境变量 - 推荐用于 MCP 环境
浏览器 cookies - 仅本地开发环境可用
# 从 Chrome 导出 cookies（需要先登录 YouTube/Bilibili）
# 使用具体视频 URL 比首页更快（避免解析推荐列表）
yt-dlp --cookies-from-browser chrome --cookies ~/.yt-cookies.txt \
--skip-download "https://www.youtube.com/watch?v=jNQXAC9IVRw"
# 设置文件权限（保护敏感信息）
chmod 600 ~/.yt-cookies.txt⚠️ 注意: Cookies 会过期（通常 1-2 周），需要定期重新导出。
# 在 shell 配置文件中添加 (~/.zshrc 或 ~/.bashrc)
export YT_DLP_COOKIES=~/.yt-cookies.txt{
"mcpServers": {
"subtitle-analyzer": {
"command": "python",
"args": ["-m", "src.server"],
"cwd": "/path/to/subtitle-analyzer-mcp",
"env": {
"YT_DLP_COOKIES": "/Users/yourname/.yt-cookies.txt"
}
}
}
}# 注册时通过 env 命令注入环境变量
claude mcp add subtitle-analyzer -- \
env YT_DLP_COOKIES=/path/to/.yt-cookies.txt \
python -m src.server --cwd /path/to/subtitle-analyzer-mcp在调用工具时直接传入 cookies_file 参数：
{
"url": "https://www.youtube.com/watch?v=xxxxx",
"cookies_file": "/path/to/.yt-cookies.txt"
}请提取这个视频的字幕：https://www.youtube.com/watch?v=xxxxx在这个视频中搜索"机器学习"出现的位置：https://www.bilibili.com/video/BVxxxxx获取这个视频的基本信息：https://www.youtube.com/watch?v=xxxxx提取这个视频的字幕并生成摘要：https://www.youtube.com/watch?v=xxxxx工具名称 | 功能 | 参数 |
| 提取字幕 | url, lang?, format?, cookies_file? |
| 搜索关键词时间戳 | url, keywords, context_lines?, cookies_file? |
| 获取视频信息 | url, cookies_file? |
| 列出可用字幕 | url, cookies_file? |
带
?的参数为可选参数
A: MCP 服务器运行在后台进程，无法直接访问浏览器 cookies。解决方法：
导出 cookies 文件：
yt-dlp --cookies-from-browser chrome --cookies ~/.yt-cookies.txt \
--skip-download "https://www.youtube.com"配置 YT_DLP_COOKIES 环境变量（见认证配置章节）
A: Bilibili 部分视频需要登录才能访问字幕。请参考认证配置章节设置 cookies。
A: 可能原因：
视频本身没有字幕 - 使用 list_available_subtitles 工具检查
需要登录认证 - 配置 cookies 文件
网络连接问题
Cookies 已过期 - 重新导出 cookies 文件
A: 自动生成的字幕（ASR）质量取决于平台算法。建议：
优先使用手动上传的字幕
使用 list_available_subtitles 查看可用选项
A: 通常 1-2 周后 cookies 会过期，届时需要重新执行导出命令。如果突然出现认证失败，请先尝试重新导出 cookies。
# 安装开发依赖
pip install -e ".[dev]"
# 运行测试
pytest
# 本地测试服务器
python -m src.serverMIT License
从视频 URL 提取字幕。 支持平台：YouTube、Bilibili 返回：完整字幕文本，包含时间戳
认证配置（按优先级）：
cookies_file 参数
YT_DLP_COOKIES 环境变量
浏览器 cookies（仅本地环境）
参数：
url: 视频链接 (必填)
lang: 首选字幕语言，如 'zh', 'en', 'ja' (可选，默认自动检测)
format: 输出格式 'text' (纯文本) 或 'srt' (带时间戳) (可选，默认 'srt')
cookies_file: cookies 文件路径，用于认证需要登录的视频 (可选)
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | 视频 URL (YouTube 或 Bilibili) | |
| lang | No | 首选字幕语言代码，如 zh, en, ja | zh |
| format | No | 输出格式：text (纯文本) 或 srt (带时间戳) | srt |
| cookies_file | No | cookies 文件路径，用于认证需要登录的视频 |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the full burden. It discloses supported platforms, return type (complete subtitle text with timestamps), and authentication fallback hierarchy. It does not mention error handling or rate limits, but sufficiently covers expected behavior for a read operation.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is well-structured with clear sections (purpose, platforms, return, authentication, parameters), front-loaded with the main action, and no extraneous information.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no output schema, the description adequately explains return values and formatting. It covers input parameters, authentication, and supported platforms. Lacks edge-case handling but is complete for typical use.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Despite 100% schema coverage, the description adds value by explicitly stating defaults (lang: auto-detect, format: 'srt') and providing authentication priority context for 'cookies_file'. This goes beyond the schema's parameter descriptions.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states '从视频 URL 提取字幕' (extract subtitles from video URL), specifying supported platforms (YouTube, Bilibili) and distinguishing this tool from siblings like 'list_available_subtitles' and 'search_timestamp' by focusing on actual extraction.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides authentication configuration priorities, offering some usage context, but does not explicitly state when to use this tool versus siblings (e.g., when to use 'list_available_subtitles' instead). No direct comparison to alternatives.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
获取视频的基本信息，包括标题、时长、描述等。
认证配置（按优先级）：
cookies_file 参数
YT_DLP_COOKIES 环境变量
浏览器 cookies（仅本地环境）
参数：
url: 视频链接 (必填)
cookies_file: cookies 文件路径，用于认证需要登录的视频 (可选)
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | 视频 URL | |
| cookies_file | No | cookies 文件路径，用于认证需要登录的视频 |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries full burden. It discloses authentication behavior (priority order) but does not mention other traits like read-only nature, idempotency, or rate limits.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise with two paragraphs: purpose and auth priority. It is front-loaded but could be slightly more streamlined.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
With no output schema, the description lists returned fields (title, duration, description). It covers both parameters and provides sibling context. Missing error handling details.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so baseline is 3. The description repeats parameter info and adds authentication priority context for cookies_file, but adds little beyond the schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool retrieves basic video information (title, duration, description). It distinguishes from sibling tools focused on subtitles and timestamp search.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides authentication configuration priority, guiding the agent on when to use each method. However, it lacks explicit when-not-to-use or comparisons with alternatives.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
列出视频可用的所有字幕语言。 用于了解视频有哪些字幕可以提取。
认证配置（按优先级）：
cookies_file 参数
YT_DLP_COOKIES 环境变量
浏览器 cookies（仅本地环境）
参数：
url: 视频链接 (必填)
cookies_file: cookies 文件路径，用于认证需要登录的视频 (可选)
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | 视频 URL | |
| cookies_file | No | cookies 文件路径，用于认证需要登录的视频 |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description discloses authentication behavior and required parameters. It does not mention other behavioral traits such as read-only nature, error handling, or rate limits. Adequate but not comprehensive.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise and well-structured, with a clear purpose statement followed by authentication configurations and parameter details. No unnecessary information is present.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's simplicity (2 parameters, no output schema), the description covers the essential aspects: purpose, authentication, and parameter details. It lacks a description of the return format, but the purpose implies a list of languages, which is sufficient.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so baseline is 3. The description adds minimal value beyond the schema, essentially repeating the parameter names and brief descriptions. No additional constraints or formatting details are provided.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's function: listing available subtitle languages for a video. It explicitly distinguishes from sibling tools like extract_subtitles by indicating that this tool is for discovering which subtitles are available.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies the primary use case (knowing which subtitles can be extracted) and provides authentication configuration guidance. However, it does not explicitly state when to use this tool versus alternatives like extract_subtitles or get_video_info.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
在字幕中搜索关键词，返回匹配的时间戳位置。 用于快速定位视频中提到某个话题的时间点。
认证配置（按优先级）：
cookies_file 参数
YT_DLP_COOKIES 环境变量
浏览器 cookies（仅本地环境）
参数：
url: 视频链接 (必填)
keywords: 要搜索的关键词列表 (必填)
context_lines: 返回匹配位置前后的上下文行数 (可选，默认 2)
cookies_file: cookies 文件路径，用于认证需要登录的视频 (可选)
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | 视频 URL | |
| keywords | Yes | 要搜索的关键词列表 | |
| cookies_file | No | cookies 文件路径，用于认证需要登录的视频 | |
| context_lines | No | 返回匹配位置前后的上下文行数 |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description must provide behavioral details. It covers authentication methods (cookies_file, env variable, browser cookies) but does not disclose potential errors (e.g., no subtitles found, unsupported video) or output format. This is adequate but not fully comprehensive.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise, with a clear two-line purpose, followed by structured sections for authentication and parameters. Every sentence adds value without redundancy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a tool with 4 parameters, no output schema, and no annotations, the description explains purpose, authentication, and parameters adequately. However, it omits the return format (e.g., timestamp list with context) and error handling, leaving some gaps for an LLM agent.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the schema already documents parameters. The description's parameter list adds minimal extra value beyond restating purposes. The authentication section provides additional context, but baseline 3 is appropriate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool searches keywords in subtitles and returns timestamp positions, with a usage context of quickly locating topics in a video. This distinguishes it from sibling tools like extract_subtitles (extracts full subtitles) and list_available_subtitles (lists available subtitle tracks).
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description includes a usage context ('用于快速定位视频中提到某个话题的时间点') but does not explicitly state when not to use this tool or mention alternatives. For example, it could clarify that subtitles must be available first (use list_available_subtitles) or that extract_subtitles is for obtaining full text.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v0.1.0extract_subtitlesget_video_infolist_available_subtitlessearch_timestampScored across 4 tools
The extract_subtitles tool has a 'get_video_info' parameter that overlaps with the separate get_video_info tool, causing ambiguity. Additionally, extract_subtitles and search_timestamp both involve subtitle content but with different purposes, though this is less problematic.
All tool names use consistent snake_case and follow a verb_noun pattern (extract_subtitles, get_video_info, list_available_subtitles, search_timestamp), making them predictable.
With 4 tools covering extraction, listing, info retrieval, and searching, the count is appropriate for the server's subtitle analysis purpose, neither too many nor too few.
The set covers core subtitle analysis operations, but missing features like subtitle translation or editing are minor gaps given the 'analyzer' focus. The inclusion of format options and search is good.
An MCP server that gives any LLM or agent clean YouTube transcripts on demand: a single video, a whole channel, or a playlist, plus AI cleanup of auto-generated captions. API-key auth, credit-based, same backend as the public v1 API. Get a free API key with 25 free credits at youtubetranscriptdownload.com/account.
YouTube transcripts, search, channel browsing, and playlists for AI agents via MCP.
An MCP server that provides tools to discover and retrieve podcast episodes transcripts.
MCP server for Hailuo (MiniMax) AI video generation