WebSearch

WebSearch is an MCP server that provides web search, targeted data scraping, and crawling capabilities by integrating the Firecrawl API with Model Context Protocol clients. Built in Python using the uv package manager, it enables LLM interfaces to execute search queries, crawl web domains to specified depths, and extract clean information using structured natural language instructions. Developers, researchers, and automated agent workflows use WebSearch to retrieve fresh online data without manually navigating web browsers or dealing with anti-scraping protections. The server connects directly to Firecrawl endpoints and supports optional integrations with OpenAI and Tavily to enhance parsing and discovery. Operating as a standard MCP toolset, it converts remote web pages into structured JSON, HTML, or Markdown text, allowing AI assistants to analyze documentation, monitor competitor domains, read online content, and supply grounded factual context across conversations and automated tasks.

Category: Browser & Web Automation

Tags: content extraction, firecrawl, search, web scraping

Visit WebSearch

How to install and configure WebSearch

1. Clone the repository: bash git clone https://github.com/josemartinrodriguezmortaloni/webSearch-Tools.git cd webSearch-Tools 2. Create a virtual environment and sync dependencies using uv: bash uv venv uv sync 3. Create a .env file in the project root and add your API keys: env FIRECRAWL_API_KEY=your_firecrawl_api_key OPENAI_API_KEY=your_openai_api_key 4. Open your Claude for Desktop configuration file and add the server to mcpServers: json { "mcpServers": { "websearch": { "command": "uv", "args": [ "--directory", "/ABSOLUTE/PATH/TO/webSearch-Tools", "run", "main.py" ] } } } 5. Replace the path with your project location, save the file, and restart Claude for Desktop.

What you can do with WebSearch

• Running targeted web searches to retrieve structured JSON query results directly inside client chat interfaces. - Extracting specific textual data from multiple URLs simultaneously using natural language guidance and optional web search. - Crawling entire website hierarchies up to defined depth limits to collect documentation in Markdown or HTML. - Scraping single web pages to extract clean page contents and optional screenshots for immediate analysis.

Key facts

• https://github.com/josemartinrodriguezmortaloni/webSearch-Tools
• Browser & Web Automation, Web Search & Research
• content extraction, firecrawl, search, web scraping

Part of MCP Servers

How do I install WebSearch?

You can install WebSearch by cloning its repository from GitHub, setting up a Python virtual environment, and installing dependencies with the uv package manager. Configure your Firecrawl and OpenAI API keys in a local .env file, then add the execution command pointing to main.py inside your Claude for Desktop configuration file.

What can WebSearch do?

WebSearch provides four core operations: running web searches via API, extracting targeted information from specified URLs using natural language instructions, crawling domains recursively with configurable depth limits, and scraping single web page contents into clean formats such as Markdown or HTML.

Which MCP clients work with WebSearch?

WebSearch works with any client supporting the standard Model Context Protocol, including Claude for Desktop and Cursor. Users configure it as a local stdio process executed through uv, allowing client tools menus to trigger web crawling, scraping, and search functions directly.

Is WebSearch open source?

Yes, WebSearch is open source software distributed under the MIT License. The code is publicly accessible on GitHub, allowing developers to inspect the source, submit feature pull requests, and customize server behavior for their own automation needs.