BACH YouTube API MCP Server
Enables access to YouTube data including channel videos, live streams, shorts, video details, transcripts, screenshots, and related content through the Yt API with support for localization and pagination.
README
Yt Api MCP Server
用于访问 Yt Api API 的 MCP 服务器。
🚀 使用 EMCP 平台快速体验
EMCP 是一个强大的 MCP 服务器管理平台,让您无需手动配置即可快速使用各种 MCP 服务器!
快速开始:
- 🌐 访问 EMCP 平台
- 📝 注册并登录账号
- 🎯 进入 MCP 广场,浏览所有可用的 MCP 服务器
- 🔍 搜索或找到本服务器(
bach-yt_api) - 🎉 点击 "安装 MCP" 按钮
- ✅ 完成!即可在您的应用中使用
EMCP 平台优势:
- ✨ 零配置:无需手动编辑配置文件
- 🎨 可视化管理:图形界面轻松管理所有 MCP 服务器
- 🔐 安全可靠:统一管理 API 密钥和认证信息
- 🚀 一键安装:MCP 广场提供丰富的服务器选择
- 📊 使用统计:实时查看服务调用情况
立即访问 EMCP 平台 开始您的 MCP 之旅!
简介
这是一个 MCP 服务器,用于访问 Yt Api API。
- PyPI 包名:
bach-yt_api - 版本: 1.0.0
- 传输协议: stdio
安装
从 PyPI 安装:
pip install bach-yt_api
从源码安装:
pip install -e .
运行
方式 1: 使用 uvx(推荐,无需安装)
# 运行(uvx 会自动安装并运行)
uvx --from bach-yt_api bach_yt_api
# 或指定版本
uvx --from bach-yt_api@latest bach_yt_api
方式 2: 直接运行(开发模式)
python server.py
方式 3: 安装后作为命令运行
# 安装
pip install bach-yt_api
# 运行(命令名使用下划线)
bach_yt_api
配置
API 认证
此 API 需要认证。请设置环境变量:
export API_KEY="your_api_key_here"
环境变量
| 变量名 | 说明 | 必需 |
|---|---|---|
API_KEY |
API 密钥 | 是 |
PORT |
不适用 | 否 |
HOST |
不适用 | 否 |
在 Cursor 中使用
编辑 Cursor MCP 配置文件 ~/.cursor/mcp.json:
{
"mcpServers": {
"bach-yt_api": {
"command": "uvx",
"args": ["--from", "bach-yt_api", "bach_yt_api"],
"env": {
"API_KEY": "your_api_key_here"
}
}
}
}
在 Claude Desktop 中使用
编辑 Claude Desktop 配置文件 claude_desktop_config.json:
{
"mcpServers": {
"bach-yt_api": {
"command": "uvx",
"args": ["--from", "bach-yt_api", "bach_yt_api"],
"env": {
"API_KEY": "your_api_key_here"
}
}
}
}
可用工具
此服务器提供以下工具:
channel_videos_via_post
Get videos listing along with channel details Quota cost is 2 unit.
端点: POST /channel/videos
参数:
-
id(string): Example value: UCMhe8jmac-tWKU0Vj_OXDvA -
forUsername(string): If id not available then provide username. Quota cost +1. -
sort_by(string): Sorts the channel videos. Available options: newest [default] oldest popular -
token(string): Pagination token -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
local(string): Get localized data, the value must be 1. Works with provided value of geo, if not provided then default is US. Quota Cost +1
channel_videos
Get channel video listing and channel details. Quota cost is 1 unit.
端点: GET /channel/videos
参数:
-
id(string): Channel Id -
forUsername(string): If id not available then provide username. Quota cost +1. -
sort_by(string): Sorts the channel videos. Available options: newest [default] popular oldest -
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Locale/language for request. Like en, gb, hi, etc -
local(string): Get localized data, the value must be 1. Works with provided value of geo, if not provided then default is US. Quota Cost +1 -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
channel_livestreams_via_post
Get livestreams listing along with channel details Quota cost is 2 unit.
端点: POST /channel/liveStreams
参数:
-
id(string): Example value: UCAuUUnT6oDeKwE6v1NGQxug -
forUsername(string): If id not available then provide username. Quota cost +1. -
sort_by(string): Sorts the channel videos. Available options: newest [default] oldest popular -
token(string): Pagination token -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
local(string): Get localized data, the value must be 1. Works with provided value of geo, if not provided then default is US. Quota Cost +1
channel_livestreams
Get live stream listing and channel details. Quota cost is 1 unit.
端点: GET /channel/liveStreams
参数:
-
id(string): Example value: UCAuUUnT6oDeKwE6v1NGQxug -
forUsername(string): If id not available then provide username. Quota cost +1. -
sort_by(string): Sorts the channel's live streams. Available options: newest [default] popular oldest -
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
local(string): Get localized data, the value must be 1. Works with provided value of geo, if not provided then default is US. Quota Cost +1 -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
channel_shorts_via_post
Get shorts listing along with channel details Quota cost is 2 unit.
端点: POST /channel/shorts
参数:
-
id(string): Example value: UCMhe8jmac-tWKU0Vj_OXDvA -
forUsername(string): If id not available then provide username. Quota cost +1. -
sort_by(string): Sorts the channel videos. Available options: newest [default] oldest popular -
token(string): Pagination token -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
local(string): Get localized data, the value must be 1. Works with provided value of geo, if not provided then default is US. Quota Cost +1
channel_shorts
Get shorts listing along with channel details Quota cost is 1 unit.
端点: GET /channel/shorts
参数:
-
id(string): Example value: UCMhe8jmac-tWKU0Vj_OXDvA -
forUsername(string): If id not available then provide username. Quota cost +1. -
sort_by(string): Sorts the channel videos. Available options: newest [default] oldest popular -
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
local(string): Get localized data, the value must be 1. Works with provided value of geo, if not provided then default is US. Quota Cost +1 -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
channel_home
Get Channel's Home Tab Listings. Quota cost is 1 unit.
端点: GET /channel/home
参数:
-
id(string): Channel Id -
forUsername(string): If id not available then provide username. Quota cost +1. -
token(string): Pagination token Alternatively, use X-TOKEN in header -
params(string): Example value: -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
screenshot
Get screenshot in JPG format. Quota Cost +4
端点: GET /video/screenshot
参数:
-
id(string) 必需: Youtube video id -
timeStamp(string) 必需: Format must be HH:MM:SS
related_videos_or_playlists
Get related videos or playlists. Quota cost is 1 unit.
端点: GET /related
参数:
-
id(string) 必需: Video id -
token(string): Pagination token -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Locale/language for request. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
metadata_update
Update viewCount, likeCount. Useful for Live videos. Quota Cost +1
端点: GET /updated_metadata
参数:
id(string) 必需: Youtube video id or multiple ids separated by , For mutli id batch request, quota cost is +1 for each extra id. Note: Mutli id batch request is experimental.
transcript
Get Transcript. Quota Cost +1
端点: GET /get_transcript
参数:
-
id(string) 必需: Youtube video id -
params(string): Example value: -
lang(string): Language code for localized results. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1
video_details
Get all the details of the YouTube video. Quota cost is 1 unit.
端点: GET /video/info
参数:
-
id(string) 必需: Youtube video id or multiple ids separated by , For mutli id batch request, quota cost is +1 for each extra id. Note: Mutli id batch request is experimental. -
extend(string): Extend the results to include more details like related videos/playlist, comment count, etc. Available options: 1 = for likeCount, commentCount, subscriberCountText, relatedVideos, channelBadges, chapters, merchandise,extraMeta,videoAttribution (not supported for multi ids) 2 = for likeCount, commentCount, channelHandle Quota cost is +1 -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1
converttranslatedownload_subtitle
Convert, translate, download the subtitle. To get the subtitle url, use Video -\u003e Subtitles endpoint or Video -\u003e Details endpoint Quota cost is 1 unit.
端点: GET /subtitle
参数:
-
url(string) 必需: Provide the Subtitle url. Available in the Video -> Subtitles or Video -> Details endpoint response. -
format(string): Subtitle format options: json3 [mime: json] srv1 [mime: xml] [default] srv2 [mime: xml] srv3 [mime: xml] ttml [mime: xml] vtt [mime: text] srt [mime: text] -
targetLang(string): Translate to Language. Valid codes are in the translationLanguages param of the Video -> Subtitles or Video -> Info endpoint response. Or provide ISO-639 code of the language like es, zh-Hans, co, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1
subtitles
Get the list of available subtitles for the video. Quota cost is 1 unit.
端点: GET /subtitles
参数:
-
id(string) 必需: Example value: arj7oStGLkU -
format(string): Subtitle format options: json3 [mime: json] srv1 [mime: xml] [default] srv2 [mime: xml] srv3 [mime: xml] ttml [mime: xml] vtt [mime: text] srt [mime: text] -
X-CACHEBYPASS(string): Example value:
downloadstream
Stream or download info. Quota cost is 1 units.
端点: GET /dl
参数:
-
id(string) 必需: Youtube Video Id or Shorts Id. -
cgeo(string) 必需: Country code in ISO 3166 format of the end user. Not providing cgeo param may cost +1 quota. It is important to provide geo of the end user to get the best speed and direct links. If links are used in the server, then cgeo will be the geo of the server. Not providing cgeo param may lead to 403 issue -
lang(string): Language code for audio results based on availability. Like en, ru, hi, etc
channel_store
Get store product listing and channel details. Quota cost is 1 unit.
端点: GET /channel/store
参数:
-
id(string): Channel Id -
forUsername(string): If id not available then provide username. Quota cost +1. -
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
query_suggestion__auto_completion
Query Suggestion / Auto-Completion.
端点: GET /suggest_queries
参数:
-
query(string) 必需: Example value: cat -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Locale/language for request. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1
about_channel
Get channel details. Quota cost is 2 unit.
端点: GET /channel/about
参数:
-
id(string) 必需: Channel Id -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
forUsername(string): Example value: -
lang(string): Language code for localized results. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1
post_details
Get post details. Quota cost is 1 unit.
端点: GET /post/info
参数:
-
id(string) 必需: Post Id -
channelId(string): Channel Id -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Locale/language for request. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1
shorts_sequence
Get Shorts sequence. Quota cost is 2 units.
端点: GET /shorts/sequence
参数:
-
params(string): Sequence param Provide either of these values: sequenceContiuation value from Shorts Info Endpoint's response for WHATTOWATCH continuation value from the previous request's response -
id(string): If the params value is not available then Shorts Video Id may be provided. But it is not recommended. -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1
shorts_details
Get Youtube Shorts video info or get Suggested Shorts video info. Quota cost is 1 unit for Shorts Video Info Quota cost is 2 units for suggested Shorts Video Info.
端点: GET /shorts/info
参数:
-
id(string) 必需: Available options: Shorts Video Id to get info or WHATTOWATCH for suggested Shorts along with info. -
params(string): Shorts video param -
extend(string): Extend the results to include more details. The value must be 1 Quota cost is +1 -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1
trending
Get trending videos region-wise \u0026 niche-wise. Quota cost is 1 unit.
端点: GET /trending
参数:
-
geo(string) 必需: ISO 3166-2 country code of the region for which you want the trending data. Like US (default), GB, CA, IN, etc. -
type(string): Trending type: now music games movies Default is now -
lang(string): Locale/language for request. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1
search
Search videos, playlists, channels, or all of them. Quota cost is 2 units.
端点: GET /search
参数:
-
query(string) 必需: Search term -
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Locale/language for request. Like en, gb, hi, etc -
type(string): Search type filter options: video channel playlist movie show shorts Sorting is not available in shorts -
duration(string): Duration filter options: short - less than 4 min medium - 4 to 20 min long - more than 20 min -
features(string): Video Features options: HD subtitles CCommons 3D Live Purchased 4K 360 Location HDR VR180 Multiple features could be joined by ',' For example: HD,subtitles -
upload_date(string): Upload Date filter options: hour today week month year -
sort_by(string): Results Sort options: relevance [default] rating date views -
local(number): Get localized data, the value must be 1. Works with provided value of geo, if not provided then default is US. Quota Cost +1 -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
comments
Get video's or shorts' comments/thread. Quota cost is 1 unit.
端点: GET /comments
参数:
-
id(string) 必需: Video Id or Shorts Video Id -
token(string): Pagination token Alternatively, use X-TOKEN in header -
sort_by(string): Available options: newest top [default] -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Locale/language for request. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
playlist
Get playlist details and video listing. Quota cost is 1 unit.
端点: GET /playlist
参数:
-
id(string) 必需: Playlist Id -
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Locale/language for request. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
home_feed
Get home feeds, region-wise \u0026 niche-wise. Quota cost is 1 unit.
端点: GET /home
参数:
-
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Locale/language for request. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
hashtag
Get videos or Shorts listings related to any hashtag. Quota cost is 1 unit.
端点: GET /hashtag
参数:
-
tag(string) 必需: Example value: viral -
type(string): Available options are all for videos and shorts shorts for only shorts -
params(string): Hashtag params -
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
post_comments
Get post's comments or thread. Quota cost is 1 unit.
端点: GET /post/comments
参数:
-
id(string) 必需: Post Id -
channelId(string): Channel Id -
sort_by(string): Available options: newest top [default] -
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Locale/language for request. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
shorts_sound_attribution
Get the Shorts original sound attribution listing. Quota cost is 1 unit.
端点: GET /shorts/soundAttribution
参数:
-
params(string): Attribution param available in the Shorts info Endpoint: soundAttribution -> params -
id(string): If the params value is not available then Shorts Video Id may be provided. -
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
search_channels_videosplaylists
Search channel's videos and playlists, also get channel details. Quota cost is 1 unit.
端点: GET /channel/search
参数:
-
id(string): Channel Id -
forUsername(string): If id not available then provide username. Quota cost +1. -
query(string) 必需: Search query -
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
channel_featured_channels
Get featured channel listing and channel details. Quota cost is 1 unit.
端点: GET /channel/channels
参数:
-
id(string): Channel Id -
forUsername(string): If id not available then provide username. Quota cost +1. -
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
channel_playlists
Get playlist listing and channel details. Quota cost is 1 unit.
端点: GET /channel/playlists
参数:
-
id(string) 必需: Example value: UCMhe8jmac-tWKU0Vj_OXDvA -
sort_by(string): Sorts the channel playlists. Available options: date_added [default] last_video_added -
token(string): Pagination token Alternatively, use X-TOKEN in header -
geo(string): ISO 3166-2 country code of the region. Like US (default), GB, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
forUsername(string): Example value: -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-TOKEN(string): Alternative way to pass Pagination token via headers. This will solve 414, URI too long issue.
resolve_url
Resolve URLs like handles, etc. Quota cost is 1 unit.
端点: GET /resolve
参数:
-
url(string) 必需: Example value: https://www.youtube.com/@TED -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1
channel_community_posts
Get community post listing and channel details. Quota cost is 1 unit.
端点: GET /channel/community
参数:
-
id(string) 必需: Channel Id -
token(string): Pagination Token -
geo(string): ISO 3166-2 country code of the region. Like US (default), UK, CA, IN, etc. -
lang(string): Language code for localized results. Like en, gb, hi, etc -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1 -
X-CACHEBYPASS(string): Cache bypass, the value must be 1 Quota Cost +1
技术栈
- 传输协议: stdio
- HTTP 客户端: httpx
许可证
MIT License - 详见 LICENSE 文件。
开发
此服务器由 API-to-MCP 工具生成。
版本: 1.0.0
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。