Which integrations are available for this server?

Allows searching for academic preprints on arXiv. Allows searching for Chinese video content on Bilibili. Allows searching the web via Brave Search with privacy focus. Allows searching the web via DuckDuckGo with privacy protection. Allows searching for code repositories on GitHub. Allows searching the web via Google. Allows searching for academic literature via Google Scholar. Allows searching the web via Mojeek's independent index. Allows searching for JavaScript packages on npm. Allows querying locations and points of interest on OpenStreetMap. Allows searching biomedical literature on PubMed. Allows searching for Python packages on PyPI. Allows searching for community discussions on Reddit. Allows searching the web via SearXNG meta-search engine. Allows searching for academic papers via Semantic Scholar. Allows searching the web via Startpage with privacy from Google results. Allows searching for high-quality images on Unsplash. Allows searching for videos on YouTube.

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Scout search for Python async web frameworks" 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.

Scout

by duguobao812718-wq

Overview Schema Source Related Servers Discussions Admin

Python

Remote

Scout

AI Agent 全能搜索工具

完全免费 · 27 引擎 · 多模态 · MCP 协议

快速开始 · 功能特性 · 搜索引擎 · MCP 工具 · 配置选项

🎯 为什么选择 Scout？

Scout 是专为 AI Agent 设计的全能搜索工具，通过 MCP 协议提供标准化的搜索接口。

特性	Scout	Tavily	Exa	SerpAPI
免费使用	✅ 完全免费	❌ 有限额	❌ 有限额	❌ 付费
无需 API Key	✅	❌	❌	❌
搜索引擎数量	26	1	1	1
多模态搜索	✅ 视频/图片/播客	❌	❌	❌
MCP 协议	✅ 原生支持	❌	❌	❌
本地缓存	✅ SQLite/Redis	❌	❌	❌

Related MCP server: Web Search MCP Server

✨ 功能特性

🔍 27 个搜索引擎

通用搜索 (8)    学术搜索 (4)    代码搜索 (2)
    ↓               ↓               ↓
社区搜索 (3)    包管理 (3)      知识搜索 (1)
    ↓               ↓               ↓
新闻搜索 (1)    多模态搜索 (4)   地图搜索 (1)
                    ↓               ↓
        YouTube · Bilibili ·     OpenStreetMap
        Unsplash · Podcast

🎬 多模态搜索

视频搜索 — YouTube、Bilibili 视频教程
图片搜索 — Unsplash 高质量图片
播客搜索 — Podcast 节目发现
地图搜索 — OpenStreetMap 地点查询

⚡ 高性能架构

并发搜索 — asyncio 异步并发，不阻塞
智能缓存 — SQLite + FTS5 全文搜索
结果合并 — RRF (Reciprocal Rank Fusion) 算法
熔断器 — 引擎连续失败自动暂停

🛡️ 安全防护

SSRF 防护 — DNS 解析 + IP 范围验证
并发限制 — Semaphore 控制最大并发数
重定向验证 — 每次重定向都检查安全性

🚀 快速开始

安装

# 克隆仓库
git clone https://github.com/duguobao812718-wq/scout.git
cd scout

# 安装依赖
pip install -e .

# 注册到 Claude Code
claude mcp add -s user scout -- python -m src

重启 Claude Code，即可使用 12 个搜索工具。

验证安装

# 列出可用引擎
python -m src --list-engines

# 运行测试
python -m pytest tests/ -v

🔍 搜索引擎

通用搜索

引擎	特点	适用场景
Google	最全的搜索结果	通用搜索
Bing	微软搜索引擎	通用搜索
Brave	隐私保护	隐私搜索
DuckDuckGo	无追踪	隐私搜索
Mojeek	独立索引	替代选择
Startpage	Google 隐私前端	隐私搜索
Yandex	俄罗斯搜索引擎	俄语内容
SearXNG	元搜索，多实例	自建实例

学术搜索

引擎	特点	适用场景
Semantic Scholar	AI 驱动的学术搜索	论文发现
arXiv	预印本论文	最新研究
Google Scholar	最全的学术搜索	论文引用
PubMed	生物医学文献	医学研究

代码搜索

引擎	特点	适用场景
GitHub	代码仓库搜索	开源项目
StackOverflow	编程问答	问题解决

社区搜索

引擎	特点	适用场景
Reddit	社区讨论	深度讨论
HackerNews	技术新闻	技术趋势
Twitter/X	社交媒体	实时动态

包管理搜索

引擎	特点	适用场景
npm	JS/TS 包	前端开发
PyPI	Python 包	Python 开发
HuggingFace	AI 模型/数据集	AI 开发

多模态搜索

引擎	特点	适用场景
YouTube	视频搜索	视频教程
Bilibili	中文视频	技术分享
Unsplash	图片搜索	高质量图片
Podcast	播客搜索	音频内容

地图搜索

引擎	特点	适用场景
OpenStreetMap	免费地图 API	地点查询、地址搜索、POI 搜索

🛠️ MCP 工具

工具	功能	示例
`search`	多引擎搜索	`search("Python tutorial")`
`fetch`	抓取页面	`fetch("https://python.org")`
`fetch_batch`	批量抓取	`fetch_batch([url1, url2])`
`research`	深度研究	`research("What is ML?")`
`summarize`	搜索+摘要	`summarize("latest AI news")`
`read_doc`	读取 PDF	`read_doc("paper.pdf")`
`engines`	列出引擎	`engines()`
`cache_search`	搜索缓存	`cache_search("Python")`
`semantic_search`	语义搜索	`semantic_search("machine learning")`
`extract_structured`	提取结构化数据	`extract_structured(url)`

⚙️ 配置选项

环境变量

在项目目录下创建 .env 文件：

# 代理配置（国内用户）
SCOUT_PROXY=http://127.0.0.1:7897

# 缓存后端
SCOUT_CACHE_BACKEND=sqlite  # 或 redis
SCOUT_REDIS_URL=redis://localhost:6379/0

# 速率限制
SCOUT_RATE_LIMIT_PER_MINUTE=60

# 安全搜索
SCOUT_SAFESEARCH=moderate  # off / moderate / strict

高级配置

# 请求超时
SCOUT_REQUEST_TIMEOUT=15

# 最大结果数
SCOUT_MAX_RESULTS=10

# 缓存 TTL
SCOUT_CACHE_TTL_SECONDS=604800  # 7 天

📊 性能指标

指标	值
搜索引擎数量	26
MCP 工具数量	12
测试用例数量	243
代码行数	~8500
平均搜索延迟	< 2s
缓存命中延迟	< 10ms

🏗️ 项目结构

scout/
├── src/
│   ├── server.py           # MCP 服务器入口
│   ├── config.py           # 配置管理
│   ├── cache.py            # SQLite 缓存
│   ├── cache_redis.py      # Redis 缓存
│   ├── scoring.py          # 结果评分
│   ├── suggestions.py      # 搜索建议
│   ├── summary.py          # 结果摘要
│   ├── ratelimit.py        # 限速 + 熔断器
│   ├── engines/            # 26 个搜索引擎
│   │   ├── __init__.py     # 引擎基类 + 注册表
│   │   ├── google.py       # 通用搜索
│   │   ├── youtube.py      # 视频搜索
│   │   └── ...
│   └── fetchers/           # 内容抓取
│       ├── http.py         # HTTP 抓取
│       └── documents.py    # PDF 解析
├── tests/                  # 测试用例
├── pyproject.toml          # 项目配置
└── README.md               # 本文件

🤝 贡献指南

欢迎贡献新的搜索引擎！添加新引擎只需 3 步：

# 1. 创建 src/engines/myengine.py
from . import JsonApiEngine, register_engine

class MyEngine(JsonApiEngine):
    name = "myengine"
    
    def build_url(self, query, max_results, filters=None):
        return f"https://api.example.com/search?q={query}"
    
    def parse(self, data):
        # 解析返回结果
        return [SearchResult(...)]

# 2. 注册引擎
register_engine(MyEngine())

# 3. 在 __init__.py 中导入
from . import myengine

运行测试验证：

python -m pytest tests/ -v

📚 参考项目

free-search-mcp — 免费搜索 MCP
mcp-smart-searcher — 多引擎搜索
exa-mcp-server — 语义搜索
Agent-Reach — 多平台抓取

📄 许可证

MIT License

Scout — 让 AI Agent 看得更远，搜得更广

GitHub · Issues · License

Available Tools

14 tools

cache_searchB

Read-onlyIdempotent

Full-text search across previously fetched and cached pages.

Inputs: query (required), limit, format. Returns: result

enginesA

Read-onlyIdempotent

List engine names accepted by the engines= parameter of search.

Inputs: no named parameters. Returns: result

extract_structuredA

Read-onlyIdempotent

Pull JSON-LD, OpenGraph, Twitter cards from a web page.

Inputs: url (required), format. Returns: result

fetchA

Read-onlyIdempotent

Fetch one URL and return reader-mode Markdown of the main content.

Args: url: Absolute http(s) URL. extract_code: 提取代码块（默认 False）。 extract_tables: 提取表格数据（默认 False）。 format: "markdown" or "json".

Inputs: url (required), extract_code, extract_tables, format. Returns: result

fetch_batchA

Read-onlyIdempotent

Fetch multiple URLs in parallel and return their contents.

Inputs: urls (required), format. Returns: result

image_searchC

Read-only

Search for images using SearXNG meta-search engine.

Inputs: query (required), max_results, freshness, format. Returns: result

read_docC

Read-onlyIdempotent

Read an http(s) document (PDF) into Markdown.

Inputs: source (required), start, length, format. Returns: result

researchC

Read-only

One-shot research: search the web, fetch the top results, return both.

Inputs: question (required), depth, engines, fetch, +9 more. Returns: result

searchC

Read-only

Run a multi-engine web search and return a ranked, deduplicated link list.

Args: translate_to: 翻译目标语言（例如 "zh", "en", "ja"）。需要启用 LLM。

Inputs: query (required), engines, max_results, page, +12 more. Returns: result

semantic_index_pageA

Idempotent

Fetch a page and add it to the semantic search index.

Inputs: url (required), format. Returns: result

semantic_searchB

Read-onlyIdempotent

Search through semantically indexed pages using meaning-based matching.

Inputs: query (required), top_k, format. Returns: result

summarizeC

Read-only

Search, fetch, extract key points, and assess credibility.

Inputs: question (required), depth, engines, freshness, +3 more. Returns: result

video_platformsA

Read-onlyIdempotent

List available video search platforms.

Inputs: no named parameters. Returns: result

video_searchA

Read-only

Search for videos on YouTube and Bilibili.

Args: query: Search query. platforms: Video platforms to search. Default: ["youtube", "bilibili"]. max_results: Max results per platform. freshness: Time filter. format: "markdown" or "json".

Inputs: query (required), platforms, max_results, freshness, +1 more. Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

Deploy Server

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

B3.4/5.0

Scored 2026-06-23 03:20 across 14 tools

Disambiguation4/5

Tools are mostly distinct: search vs. research overlap slightly (research does search + fetch top results), and fetch vs. read_doc both fetch but read_doc is PDF-specific. Most others clearly differentiate.

Naming Consistency3/5

Naming is inconsistent with mix of verb-first (fetch, search, summarize), noun-first (cache_search, video_search), and plain nouns (engines, research). No uniform pattern but still readable.

Tool Count5/5

14 tools cover a well-scoped domain of web research, fetching, searching, caching, semantic indexing, and summarization. Each tool has a clear purpose and justifies its existence.

Completeness4/5

Covers core research workflows: search (web, image, video), fetch, read PDF, extract metadata, cache, semantic index/search, summarize. Minor gaps like cache deletion or management, but not essential for primary use.

Resources

Related MCP Connectors

web-search
ai.groundroute
Web search for AI agents — one tool across 6 engines, routed to the cheapest + cached.
Search Browser Automation
21
https://api.groundroute.ai/mcp
Ninelayer
in.ninelayer
Agent-native search engine with live web research optimized for AI agents.
Search Agent Orchestration
2
https://mcp.ninelayer.in/mcp/
qsearchA
io.github.theYahia
Multi-engine search for AI agents. Trust scoring, local corpus, MCP-native. Self-hostable, BYOK.
Search RAG Systems
24
https://qsearch.pro/mcp
HasDataA
com.hasdata.mcp
Direct access to 60+ scraping and search tools. Extract structured data from Google (Search, Maps, Trends), Amazon, Airbnb, Social Media, and any web page directly into your AI agent.
Browser Automation RAG Systems Web Scraping
742
https://mcp.hasdata.com/mcp

Related MCP Servers

Search MCP Server
Search Version Control
ilgizar-valiullin
A
license
A
quality
B
maintenance
Enables AI agents to perform unified web searches, GitHub, and GitLab searches with caching, reranking, and fallback across multiple providers.
Updated 11 days ago (2026-11-02 13:00 UTC)
4
176 npm
20
MIT
Web Search MCP Server
Browser Automation Web Scraping Search
santhoshkanna812-spec
F
license
Not graded
quality
D
maintenance
Enables AI agents to search the web and extract content using multiple search providers, with caching, retry logic, and options for JavaScript-heavy page rendering.
Updated 4 months ago (2026-07-14 16:53 UTC)
-
Web Research MCP
Search Research & Data Browser Automation
infinit3labs
A
license
A
quality
B
maintenance
Enables AI agents to search and fetch high-quality information from multiple sources, including general web APIs, Wikipedia, arXiv, Hacker News, Stack Exchange, and Crossref, with optional pro-mode deep research and clean markdown page extraction.
Updated 2 months ago (2026-09-02 23:30 UTC)
10
MIT
web-search-server
Search Web Scraping Browser Automation
sanqiuyef
A
license
Not graded
quality
C
maintenance
Enables AI agents to run local multi-engine web searches, fetch pages through anti-bot fallback chains, and query site-specific sources such as GitHub, npm, PyPI, Docker Hub, and Twitter/X for structured results.
Updated a month ago (2026-10-18 10:50 UTC)
3
MIT

Scout

🎯 为什么选择 Scout？

✨ 功能特性

🔍 27 个搜索引擎

🎬 多模态搜索

⚡ 高性能架构

🛡️ 安全防护

🚀 快速开始

安装

验证安装

🔍 搜索引擎

通用搜索

学术搜索

代码搜索

社区搜索

包管理搜索

多模态搜索

地图搜索

🛠️ MCP 工具

⚙️ 配置选项

环境变量

高级配置

📊 性能指标

🏗️ 项目结构

🤝 贡献指南

📚 参考项目

📄 许可证

Available Tools

Tool Schema Changelog

TDQS

Resources

Related MCP Connectors

Related MCP Servers

Search MCP Server

Web Search MCP Server

Web Research MCP

web-search-server