Skip to main content
Glama
enesjakozh
artivus-labs

DuckDuckGo MCP Server

by artivus-labs
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Remote

ddg-mcp MCP 서버

DuckDuckGo 검색 API MCP - Model Context Protocol을 통해 DuckDuckGo 검색 기능을 제공하는 서버입니다.

구성 요소

프롬프트

서버는 다음과 같은 프롬프트를 제공합니다.

  • search-results-summary : DuckDuckGo 검색 결과 요약을 생성합니다.

    • 검색어에 대한 필수 "쿼리" 인수

    • 세부 수준(간략/상세)을 제어하기 위한 선택적 "스타일" 인수

도구

서버는 다음의 DuckDuckGo 검색 도구를 구현합니다.

  • ddg-text-search : DuckDuckGo를 사용하여 웹에서 텍스트 결과를 검색합니다.

    • 필수: "키워드" - 검색어 키워드

    • 선택 사항: "지역", "안전 검색", "시간 제한", "최대 결과"

  • ddg-image-search : DuckDuckGo를 사용하여 웹에서 이미지 검색

    • 필수: "키워드" - 검색어 키워드

    • 선택 사항: "지역", "안전 검색", "시간 제한", "크기", "색상", "유형 이미지", "레이아웃", "라이선스 이미지", "최대 결과"

  • ddg-news-search : DuckDuckGo를 사용하여 뉴스 기사 검색

    • 필수: "키워드" - 검색어 키워드

    • 선택 사항: "지역", "안전 검색", "시간 제한", "최대 결과"

  • ddg-video-search : DuckDuckGo를 사용하여 비디오 검색

    • 필수: "키워드" - 검색어 키워드

    • 선택 사항: "지역", "안전 검색", "시간 제한", "해상도", "기간", "라이선스_비디오", "최대_결과"

  • ddg-ai-chat : DuckDuckGo AI와 채팅

    • 필수: "키워드" - AI에 보낼 메시지 또는 질문

    • 선택 사항: "model" - 사용할 AI 모델(옵션: "gpt-4o-mini", "llama-3.3-70b", "claude-3-haiku", "o3-mini", "mistral-small-3")

Related MCP server: DuckDuckGo MCP Server

설치

필수 조건

  • 파이썬 3.9 이상

  • uv (권장) 또는 pip

PyPI에서 설치

지엑스피1

소스에서 설치

  1. 저장소를 복제합니다.

git clone https://github.com/misanthropic-ai/ddg-mcp.git
cd ddg-mcp

  1. 패키지를 설치하세요:

# Using uv
uv install -e .

# Using pip
pip install -e .

구성

필수 종속성

서버에는 duckduckgo-search 패키지가 필요한데, 이 패키지는 ddg-mcp 설치하면 자동으로 설치됩니다.

수동으로 설치해야 하는 경우:

uv install duckduckgo-search
# or
pip install duckduckgo-search

DuckDuckGo 검색 매개변수

공통 매개변수

다음 매개변수는 대부분의 검색 유형에 사용할 수 있습니다.

  • 지역 : 지역화된 결과에 대한 지역 코드(기본값: "wt-wt")

    • 예: "us-en"(미국 영어), "uk-en"(영국 영어), "ru-ru"(러시아어)

    • 더 많은 옵션은 DuckDuckGo 지역을 참조하세요.

  • safesearch : 콘텐츠 필터링 수준(기본값: "보통")

    • "on": 엄격한 필터링

    • "moderate": 중간 필터링

    • "off": 필터링 안 함

  • timelimit : 결과에 대한 시간 범위

    • "d": 마지막 날

    • "w": 지난주

    • "m": 지난달

    • "y": 작년(뉴스/영상 제공 불가)

  • max_results : 반환할 최대 결과 수(기본값: 10)

검색 연산자

검색 키워드에 다음 연산자를 사용할 수 있습니다.

  • cats dogs : 고양이 또는 개에 대한 검색 결과

  • "cats and dogs" : "고양이와 개"라는 정확한 용어에 대한 검색 결과

  • cats -dogs : 검색 결과에서 개가 더 적음

  • cats +dogs : 검색 결과에 더 많은 개가 표시됩니다.

  • cats filetype:pdf : 고양이에 대한 PDF(지원: pdf, doc(x), xls(x), ppt(x), html)

  • dogs site:example.com : example.com의 개 관련 페이지

  • cats -site:example.com : example.com을 제외한 고양이 관련 페이지

  • intitle:dogs : 페이지 제목에 "dogs"라는 단어가 포함되어 있습니다.

  • inurl:cats : 페이지 URL에 "cats"라는 단어가 포함되어 있습니다.

이미지 검색 특정 매개변수

  • 크기 : "소형", "중형", "대형", "벽지"

  • 색상 : "색상", "단색", "빨간색", "주황색", "노란색", "녹색", "파란색", "보라색", "분홍색", "갈색", "검정색", "회색", "청록색", "흰색"

  • type_image : "사진", "클립아트", "gif", "투명", "선"

  • 레이아웃 : "정사각형", "높음", "넓음"

  • license_image : "모든", "공개", "공유", "상업적으로 공유", "수정", "상업적으로 수정"

비디오 검색 특정 매개변수

  • 해상도 : "높음", "표준"

  • 지속 시간 : "짧음", "중간", "길음"

  • license_videos : "creativeCommon", "youtube"

AI 채팅 모델

  • gpt-4o-mini : OpenAI의 GPT-4o 미니 모델

  • llama-3.3-70b : 메타의 Llama 3.3 70B 모델

  • claude-3-haiku : Anthropic의 Claude 3 Haiku 모델

  • o3-mini : OpenAI의 O3 미니 모델

  • mistral-small-3 : Mistral AI의 소형 모델

빠른 시작

설치하다

클로드 데스크탑

MacOS의 경우: ~/Library/Application\ Support/Claude/claude_desktop_config.json Windows의 경우: %APPDATA%/Claude/claude_desktop_config.json

사용 예

텍스트 검색

Use the ddg-text-search tool to search for "climate change solutions"

고급 예제:

Use the ddg-text-search tool to search for "renewable energy filetype:pdf site:edu" with region "us-en", safesearch "off", timelimit "y", and max_results 20

이미지 검색

Use the ddg-image-search tool to find images of "renewable energy" with color set to "Green"

고급 예제:

Use the ddg-image-search tool to find images of "mountain landscape" with size "Large", color "Blue", type_image "photo", layout "Wide", and license_image "Public"

뉴스 검색

Use the ddg-news-search tool to find recent news about "artificial intelligence" from the last day

고급 예제:

Use the ddg-news-search tool to search for "space exploration" with region "uk-en", timelimit "w", and max_results 15

비디오 검색

Use the ddg-video-search tool to find videos about "machine learning tutorials" with duration set to "medium"

고급 예제:

Use the ddg-video-search tool to search for "cooking recipes" with resolution "high", duration "short", license_videos "creativeCommon", and max_results 10

AI 채팅

Use the ddg-ai-chat tool to ask "What are the latest developments in quantum computing?" using the claude-3-haiku model

검색 결과 요약

Use the search-results-summary prompt with query "space exploration" and style "detailed"

클로드 구성

"ddg-mcp": { "명령": "uv", "인수": [ "--디렉토리", "설치 경로/ddg-mcp", "실행", "ddg-mcp" ] },

개발

건축 및 출판

배포를 위해 패키지를 준비하려면:

  1. 종속성 동기화 및 잠금 파일 업데이트:

uv sync

  1. 패키지 배포 빌드:

uv build

이렇게 하면 dist/ 디렉토리에 소스와 휠 배포판이 생성됩니다.

  1. PyPI에 게시:

uv publish

참고: 환경 변수나 명령 플래그를 통해 PyPI 자격 증명을 설정해야 합니다.

  • 토큰: --token 또는 UV_PUBLISH_TOKEN

  • 또는 사용자 이름/비밀번호: --username / UV_PUBLISH_USERNAME 및 --password / UV_PUBLISH_PASSWORD

GitHub Actions를 사용한 자동 게시

이 저장소에는 PyPI에 자동으로 게시하기 위한 GitHub Actions 워크플로가 포함되어 있습니다. 이 워크플로는 다음과 같은 경우에 트리거됩니다.

  1. 새로운 GitHub 릴리스가 생성되었습니다.

  2. 워크플로는 GitHub Actions 인터페이스를 통해 수동으로 트리거됩니다.

자동 게시를 설정하려면:

  1. PyPI API 토큰을 생성합니다.

    • https://pypi.org/manage/account/token/ 으로 이동하세요

    • ddg-mcp 프로젝트로 범위가 제한된 새 토큰을 만듭니다.

    • 토큰 값을 복사하세요(한 번만 볼 수 있습니다)

  2. GitHub 저장소 비밀에 토큰을 추가합니다.

    • GitHub의 저장소로 이동하세요

    • 설정 > 비밀 및 변수 > 작업으로 이동합니다.

    • "새로운 저장소 비밀"을 클릭하세요

    • 이름: PYPI_API_TOKEN

    • 값: PyPI 토큰을 붙여넣으세요

    • "비밀 추가"를 클릭하세요

  3. 새 버전을 게시하려면:

    • pyproject.toml 에서 버전 번호를 업데이트합니다.

    • GitHub에서 새 릴리스를 만들거나 워크플로를 수동으로 트리거합니다.

디버깅

MCP 서버는 stdio를 통해 실행되므로 디버깅이 어려울 수 있습니다. 최상의 디버깅 환경을 위해서는 MCP Inspector 사용을 강력히 권장합니다.

다음 명령을 사용하여 npm 통해 MCP Inspector를 시작할 수 있습니다.

npx @modelcontextprotocol/inspector uv --directory /path/to/your/ddg-mcp run ddg-mcp

Inspector를 실행하면 브라우저에서 접근하여 디버깅을 시작할 수 있는 URL이 표시됩니다.

Available Tools

5 tools
ddg-ai-chatC

Chat with DuckDuckGo AI

ParametersJSON Schema
NameRequiredDescriptionDefault
keywordsYesMessage or question to send to the AI
modelNoAI model to usegpt-4o-mini

TDQS

C2.7/5.0
Behavior1/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure but offers almost none. 'Chat with DuckDuckGo AI' doesn't reveal whether this is a read-only operation, if it requires authentication, what rate limits apply, whether conversations are persistent, or what the typical response format looks like. For a chat tool with zero annotation coverage, this is a significant gap in behavioral transparency.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is extremely concise at just four words, with zero wasted language. It's front-loaded with the core functionality ('Chat with DuckDuckGo AI') and every word earns its place. This is a model of efficiency in tool description writing.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given that this is a chat tool with no annotations, no output schema, and 2 parameters, the description is insufficiently complete. It doesn't explain what kind of responses to expect, whether there are conversation contexts, what the AI's capabilities or limitations are, or any behavioral characteristics. For a tool that presumably involves AI interaction, more context about the nature of the chat would be expected.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema description coverage is 100%, so both parameters are well-documented in the schema itself. The description adds no additional parameter information beyond what's already in the schema (keywords for the message, model selection from specific AI models). This meets the baseline expectation when the schema does the heavy lifting, but doesn't provide extra context about parameter usage or constraints.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description 'Chat with DuckDuckGo AI' clearly states the verb ('Chat') and resource ('DuckDuckGo AI'), making the purpose immediately understandable. It distinguishes this tool from its siblings (image-search, news-search, text-search, video-search) by specifying it's for AI chat rather than search operations. However, it doesn't specify what kind of chat (e.g., conversational, Q&A) or the scope of the AI's capabilities.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides no guidance on when to use this tool versus its sibling tools. It doesn't mention that this is for AI-powered conversations rather than traditional search operations, nor does it suggest alternatives like using text-search for factual queries. There's no context about appropriate use cases or limitations.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

ddg-text-searchC

Search the web for text results using DuckDuckGo

ParametersJSON Schema
NameRequiredDescriptionDefault
keywordsYesSearch query keywords
regionNoRegion code (e.g., wt-wt, us-en, uk-en)wt-wt
safesearchNoSafe search levelmoderate
timelimitNoTime limit (d=day, w=week, m=month, y=year)
max_resultsNoMaximum number of results to return

TDQS

C2.9/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description carries the full burden of behavioral disclosure. It states the tool performs a web search but doesn't mention any behavioral traits such as rate limits, authentication needs, response format, or potential side effects. For a search tool with no annotation coverage, this is a significant gap in transparency.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, efficient sentence that directly states the tool's purpose without any unnecessary words. It's appropriately sized and front-loaded, making it easy for an agent to parse quickly.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's complexity (a web search with 5 parameters) and the lack of both annotations and an output schema, the description is insufficient. It doesn't explain what the tool returns, how results are structured, or any behavioral aspects, leaving critical gaps for the agent to understand the tool fully.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The description adds no parameter-specific information beyond what the input schema provides. Since schema description coverage is 100%, the baseline score is 3. The description doesn't elaborate on parameter usage, constraints, or examples, so it doesn't add value beyond the schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the action ('Search the web for text results') and the resource ('using DuckDuckGo'), which is specific and unambiguous. However, it doesn't explicitly distinguish this tool from its siblings like ddg-image-search or ddg-news-search, though the 'text results' wording implies a distinction from those other search types.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides no guidance on when to use this tool versus its siblings (ddg-ai-chat, ddg-image-search, ddg-news-search, ddg-video-search). It doesn't mention any prerequisites, alternatives, or exclusions, leaving the agent to infer usage based on the tool name alone.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

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

  1. 5 tool updates
    • First observedddg-ai-chat
    • First observedddg-image-search
    • First observedddg-news-search
    • First observedddg-text-search
    • First observedddg-video-search
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 across 5 tools

Disambiguation5/5

Every tool has a clearly distinct purpose based on media type: chat, images, news, text, and videos. There is no overlap in functionality, making it easy for an agent to select the appropriate tool for each search need.

Naming Consistency5/5

All tools follow a consistent 'ddg-[media_type]-search' pattern, with the exception of 'ddg-ai-chat' which still fits the 'ddg-[function]' convention. This uniformity makes the tool set predictable and easy to understand.

Tool Count5/5

Five tools is well-scoped for a DuckDuckGo search server, covering key search types (text, image, video, news) plus an AI chat feature. Each tool earns its place without being overwhelming or insufficient.

Completeness4/5

The tool set covers major search categories effectively, but there is a minor gap in specialized searches like maps or shopping, which are common in search engines. However, core workflows are well-supported, and agents can work around this limitation.

Maintenance

ActivityInactive
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • DuckDuckGo MCP Server

    nickclyde
    A
    license
    A
    quality
    A
    maintenance
    A Model Context Protocol (MCP) server that provides web search capabilities through DuckDuckGo, with additional features for content fetching and parsing.
    3
    15,517 PyPI
    1,510
    MIT

  • DuckDuckGo MCP Server

    zhsama
    A
    license
    B
    quality
    D
    maintenance
    A Model Context Protocol server that provides DuckDuckGo search functionality for Claude, enabling web search capabilities through a clean tool interface with rate limiting support.
    1
    1,589 npm
    87
    MIT

Appeared in Searches