Skip to main content
Glama
enesjakozh
josemartinrodriguezmortaloni

WebSearch

by josemartinrodriguezmortaloni
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Hybrid

WebSearch - 高度なWeb検索およびコンテンツ抽出ツール

ライセンスPythonバージョンファイヤークロール紫外線

Python で構築された強力な Web 検索およびコンテンツ抽出ツールで、Firecrawl API を活用して高度な Web スクレイピング、検索、コンテンツ分析機能を実現します。

🚀 機能

  • 高度なウェブ検索: カスタマイズ可能なパラメータを使用してインテリジェントなウェブ検索を実行します

  • コンテンツ抽出:自然言語プロンプトを使用してWebページから特定の情報を抽出します

  • Web クロール: 設定可能な深さと制限で Web サイトをクロールします

  • Webスクレイピング: さまざまな出力形式をサポートしたWebページをスクレイピングします

  • MCP統合:シームレスな統合を実現するモデルコンテキストプロトコル(MCP)サーバーとして構築

Related MCP server: Firecrawl MCP Server

📋 前提条件

  • Python 3.8以上

  • UVパッケージマネージャー

  • Firecrawl APIキー

  • OpenAI API キー（オプション、拡張機能用）

  • Tavily API キー（オプション、追加の検索機能用）

🛠️ インストール

  1. uvをインストールします:

# On Windows (using pip)
pip install uv

# On Unix/MacOS
curl -LsSf https://astral.sh/uv/install.sh | sh

# Add uv to PATH (Unix/MacOS)
export PATH="$HOME/.local/bin:$PATH"

# Add uv to PATH (Windows - add to Environment Variables)
# Add: %USERPROFILE%\.local\bin

  1. リポジトリをクローンします。

git clone https://github.com/yourusername/websearch.git
cd websearch

  1. uv を使用して仮想環境を作成し、アクティブ化します。

# Create virtual environment
uv venv

# Activate on Windows
.\.venv\Scripts\activate.ps1

# Activate on Unix/MacOS
source .venv/bin/activate

  1. uv を使用して依存関係をインストールします。

# Install from requirements.txt
uv sync

  1. 環境変数を設定します。

# Create .env file
touch .env

# Add your API keys
FIRECRAWL_API_KEY=your_firecrawl_api_key
OPENAI_API_KEY=your_openai_api_key

🎯 使用方法

Claude のデスクトップ版の設定

サーバーを直接実行する代わりに、Claude for Desktop が WebSearch ツールにアクセスするように構成できます。

  1. Claude for Desktop 構成ファイルを見つけるか作成します。

    • Windows: %env:AppData%\Claude\claude_desktop_config.json

    • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

  2. WebSearch サーバー構成をmcpServersセクションに追加します。

{
  "mcpServers": {
    "websearch": {
      "command": "uv",
      "args": [
        "--directory",
        "D:\\ABSOLUTE\\PATH\\TO\\WebSearch",
        "run",
        "main.py"
      ]
    }
  }
}

  1. ディレクトリ パスを、WebSearch プロジェクト フォルダーへの絶対パスに置き換えてください。

  2. 設定ファイルを保存し、Claude for Desktop を再起動します。

  3. 設定が完了すると、WebSearch ツールが Claude for Desktop のツール メニュー (ハンマー アイコン) に表示されます。

利用可能なツール

  1. 検索

  2. 情報抽出

  3. ウェブサイトをクロールする

  4. コンテンツをスクレイピングする

📚 APIリファレンス

検索

  • query (str): 検索クエリ

  • 戻り値: JSON形式の検索結果

抽出する

  • urls (List[str]): 情報を抽出するURLのリスト

  • prompt (str): 抽出の指示

  • enableWebSearch (bool): 補助的なウェブ検索を有効にする

  • showSources (bool): ソース参照を含める

  • 戻り値: 指定された形式で抽出された情報

クロール

  • url (str): 開始URL

  • maxDepth (int): 最大クロール深度

  • limit (int): クロールする最大ページ数

  • 戻り値: クロールされたコンテンツをマークダウン/HTML形式で返します

削る

  • url (str): ターゲットURL

  • 戻り値: オプションのスクリーンショットを含むスクレイピングされたコンテンツ

🔧 構成

環境変数

このツールが機能するには特定のAPIキーが必要です。テンプレートとして使用できる.env.exampleファイルをご用意しております。

  1. サンプルファイルをコピーします。

# On Unix/MacOS
cp .env.example .env

# On Windows
copy .env.example .env

  1. API キーを使用して.envファイルを編集します。

# OpenAI API key - Required for AI-powered features
OPENAI_API_KEY=your_openai_api_key_here

# Firecrawl API key - Required for web scraping and searching
FIRECRAWL_API_KEY=your_firecrawl_api_key_here

APIキーの取得

  1. OpenAI APIキー:

  2. Firecrawl APIキー:

    • Firecrawlのウェブサイトをご覧ください

    • アカウントを作成する

    • ダッシュボードに移動する

    • 新しいAPIキーを生成する

すべてが正しく設定されていれば、検索結果を含む JSON 応答が返されます。

トラブルシューティング

エラーが発生した場合:

  1. .envファイルに必要なすべての API キーが設定されていることを確認します。

  2. APIキーが有効であり、期限が切れていないことを確認します

  3. .envファイルがプロジェクトのルートディレクトリにあることを確認します。

  4. 環境変数が正しく読み込まれていることを確認してください

🤝 貢献する

  1. リポジトリをフォークする

  2. 機能ブランチを作成します（ git checkout -b feature/AmazingFeature ）

  3. 変更をコミットします（ git commit -m 'Add some AmazingFeature' ）

  4. ブランチにプッシュする ( git push origin feature/AmazingFeature )

  5. プルリクエストを開く

📝 ライセンス

このプロジェクトは MIT ライセンスに基づいてライセンスされています - 詳細についてはLICENSEファイルを参照してください。

🙏 謝辞

  • 強力なウェブスクレイピングAPIを提供するFirecrawl

  • AI機能のためのOpenAI

  • MCPプロトコル仕様のMCPコミュニティ

📬 お問い合わせ

ホセ・マルティン・ロドリゲス・モルタローニ - @m4s1t425 - jmrodriguezm13@gmail.com

PythonとFirecrawlを使って❤️で作りました

Available Tools

4 tools
crawlB

Crawls a website starting from the specified URL and extracts content from multiple pages. Args: - url: The complete URL of the web page to start crawling from - maxDepth: The maximum depth level for crawling linked pages - limit: The maximum number of pages to crawl

Returns:
- Content extracted from the crawled pages in markdown and HTML format
ParametersJSON Schema
NameRequiredDescriptionDefault
urlYes
maxDepthYes
limitYes

TDQS

B3.2/5.0
Behavior2/5

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. It states the tool crawls and extracts content, implying it performs read operations, but lacks details on permissions, rate limits, potential impacts on target sites, or error handling. For a web crawling tool with zero 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 efficiently structured: a concise opening sentence states the purpose, followed by a bulleted list for args and returns. Every sentence earns its place by delivering essential information without redundancy, making it easy to parse and front-loaded with key details.

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

Completeness3/5

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

Given the tool's complexity (web crawling with 3 parameters), no annotations, and no output schema, the description is moderately complete. It covers the basic purpose and parameters but lacks details on behavioral traits, error cases, or output format specifics beyond 'markdown and HTML format'. This is adequate for a minimal viable description but has clear gaps.

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

Parameters4/5

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

The description adds substantial meaning beyond the input schema, which has 0% description coverage. It explains each parameter's purpose: 'url' as the starting point, 'maxDepth' for crawl depth, and 'limit' for page count. This compensates well for the schema's lack of descriptions, providing clear semantics for all three parameters.

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 tool's purpose: 'Crawls a website starting from the specified URL and extracts content from multiple pages.' It specifies the verb ('crawls'), resource ('website'), and scope ('extracts content from multiple pages'), making the action clear. However, it doesn't explicitly differentiate from sibling tools like 'extract' or 'scrape', which likely have overlapping functions.

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 alternatives like 'extract' or 'scrape'. It mentions the tool's function but offers no context about prerequisites, exclusions, or comparative use cases. This leaves the agent without clear direction for tool selection among siblings.

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

extractB

Extracts specific information from a web page based on a prompt. Args: - url: The complete URL of the web page to extract information from - prompt: Instructions specifying what information to extract from the page - enabaleWebSearch: Whether to allow web searches to supplement the extraction - showSources: Whether to include source references in the response

Returns:
- Extracted information from the web page based on the prompt
ParametersJSON Schema
NameRequiredDescriptionDefault
urlYes
promptYes
enabaleWebSearchYes
showSourcesYes

TDQS

B3.1/5.0
Behavior2/5

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. It mentions that the tool extracts information and includes parameters for web search and source references, but doesn't describe what happens during extraction (e.g., rate limits, authentication needs, error conditions, or what 'extracted information' looks like). For a tool with 4 parameters and no annotations, this is insufficient behavioral context.

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

Conciseness4/5

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

The description is well-structured with a clear purpose statement followed by parameter explanations and return value description. It's appropriately sized for a 4-parameter tool, though the 'Returns' section could be more specific. Every sentence adds value, and there's no unnecessary repetition or fluff.

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

Completeness3/5

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

Given the tool's complexity (4 parameters, no annotations, no output schema), the description is moderately complete. It covers the basic purpose and parameters but lacks details about behavioral traits, error handling, and what the extracted information actually contains. Without an output schema, the return value description is vague ('Extracted information from the web page based on the prompt').

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

Parameters4/5

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

The description provides clear semantic explanations for all 4 parameters beyond what the input schema offers (which has 0% description coverage). It explains that 'url' is for the web page, 'prompt' specifies what to extract, 'enableWebSearch' allows supplemental searches, and 'showSources' includes references. This adds significant value over the bare schema, though it doesn't detail parameter interactions 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 clearly states the tool's purpose: extracting specific information from a web page based on a prompt. It specifies the verb ('extracts') and resource ('web page'), but doesn't explicitly differentiate from sibling tools like 'crawl', 'scrape', or 'search' beyond the extraction focus. The description is specific about the action but lacks sibling tool comparison.

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 alternatives like 'crawl', 'scrape', or 'search'. It doesn't mention prerequisites, use cases, or exclusions. The only implied usage is for extracting information from web pages, but with no context about when this is preferable to other tools on the server.

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

scrapeD
ParametersJSON Schema
NameRequiredDescriptionDefault
urlYes

TDQS

D1/5.0
Behavior1/5

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

Tool has no description.

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

Conciseness1/5

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

Tool has no description.

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

Completeness1/5

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

Tool has no description.

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

Parameters1/5

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

Tool has no description.

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

Purpose1/5

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

Tool has no description.

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

Usage Guidelines1/5

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

Tool has no description.

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. 4 tool updatesv1.0.0
    • First observedcrawl
    • First observedextract
    • First observedscrape
    • First observedsearch
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

C2.4/5.0

Scored across 4 tools

Disambiguation2/5

The tools have significant overlap and unclear boundaries. 'crawl' extracts content from multiple pages, 'extract' pulls specific info from a single page, and 'scrape' (with no description) is ambiguous—likely overlapping with both. 'search' is distinct for web searches, but the others could easily be confused for similar web content tasks.

Naming Consistency5/5

All tool names follow a consistent, simple verb pattern (crawl, extract, scrape, search). They are short, clear, and uniformly styled without mixing conventions, making them predictable and easy to parse.

Tool Count4/5

Four tools is reasonable for a web search domain, allowing coverage of crawling, extraction, scraping, and searching. It's slightly thin but manageable, as each tool addresses a core aspect of web data retrieval without being overly bloated.

Completeness3/5

There are notable gaps in the tool surface. The server covers basic retrieval (crawl, search) and extraction, but lacks update/delete operations (e.g., no tool to modify or clear cached data) and has a dead tool ('scrape' with no description), which limits functionality. However, agents can work around this for common web search tasks.

Resources

Related MCP Connectors

Related MCP Servers

  • Firecrawl MCP Server

    ampcome-mcps
    A
    license
    A
    quality
    D
    maintenance
    A Model Context Protocol server that enables web scraping, crawling, and content extraction capabilities through integration with Firecrawl.
    8
    22,552 npm
    2
    MIT

  • MCP Firecrawl

    mustafamemon265789-cloud
    F
    license
    A
    quality
    D
    maintenance
    A production-ready Model Context Protocol (MCP) server that integrates with the Firecrawl API to give AI assistants the power to scrape, crawl, and search the web.
    3
    -

Appeared in Searches