What features does Web Search MCP MCP server provide?

Q: What features does Web Search MCP MCP server provide?

full-web-search (Main Tool): Comprehensive web search that queries Bing, Brave, and DuckDuckGo, fetches full page content from results, and returns structured results with concurrent processing and robust error handling.. get-web-search-summaries (Lightweight Alternative): Uses the same multi-engine search strategy but returns only result snippets/descriptions without full content extraction for faster results.. get-single-web-page-content (Utility Tool): Fetches and extracts the main content from a single provided URL, removing navigation, ads, and other non-content elements.. Optimised multi-engine search strategy: Prioritises Bing > Brave > DuckDuckGo with dedicated browser isolation per engine and a fallback sequence to maximise reliability and performance.. Dedicated browser isolation per engine: Each search engine runs in its own isolated browser instance with automatic cleanup to prevent cross-engine interference and resource leakage.. Concurrent content extraction: Content is extracted from multiple result pages concurrently with timeout protections to speed up results.. HTTP/2 error recovery: Automatically falls back to HTTP/1.1 when protocol errors occur to improve resilience.. Content extraction strategy and quality controls: Axios-first content extraction with browser fallback, plus configurable content length, timeouts, and quality settings via environment variables.

Web Search MCP MCP Server 2025: Features & Setup Guide

Overview

This TypeScript MCP server provides comprehensive web search capabilities without relying on API keys, offering multiple tools tailored for different use cases. It exposes three MCP tools: full-web-search (the main tool) which employs an optimised, multi-engine strategy across Bing, Brave, and DuckDuckGo with dedicated browser instances, concurrent content extraction, and fallbacks for HTTP/2 errors; get-web-search-summaries (lightweight) which returns only search result snippets and descriptions; and get-single-web-page-content (utility) which fetches and cleans a single page's main content. The search workflow prioritises browser-based Bing (Chromium) and Brave (Firefox) before using an Axios-backed DuckDuckGo search, with each engine isolated in its own browser and automatically cleaned up. Content extraction combines axios-first attempts with browser-based extraction when needed, and supports concurrent processing with timeout protection. The server automatically recovers from HTTP/2 errors by falling back to HTTP/1.1. It’s compatible with LM Studio and LibreChat, and uses environment/config settings to tune content length, timeouts, browser behavior, and search quality. Model compatibility guidance emphasizes newer tool-using models (Qwen3, Gemma3, Llama 3.x) for best results.

Features

full-web-search (Main Tool)

Comprehensive web search that queries Bing, Brave, and DuckDuckGo, fetches full page content from results, and returns structured results with concurrent processing and robust error handling.

get-web-search-summaries (Lightweight Alternative)

Uses the same multi-engine search strategy but returns only result snippets/descriptions without full content extraction for faster results.

get-single-web-page-content (Utility Tool)

Fetches and extracts the main content from a single provided URL, removing navigation, ads, and other non-content elements.

Optimised multi-engine search strategy

Prioritises Bing > Brave > DuckDuckGo with dedicated browser isolation per engine and a fallback sequence to maximise reliability and performance.

Dedicated browser isolation per engine

Each search engine runs in its own isolated browser instance with automatic cleanup to prevent cross-engine interference and resource leakage.

Concurrent content extraction

Content is extracted from multiple result pages concurrently with timeout protections to speed up results.

HTTP/2 error recovery

Automatically falls back to HTTP/1.1 when protocol errors occur to improve resilience.

Content extraction strategy and quality controls

Axios-first content extraction with browser fallback, plus configurable content length, timeouts, and quality settings via environment variables.