MixSearch

You can run and use the MCP MixSearch server to perform multi-engine web searches, extract content, and filter results with ease. It offers both a protocol-based interface and a REST API, ensuring reliable search capabilities even when some engines face issues. This guide shows you how to install, run, and configure the server, and how to perform common search tasks.

How to use

To search with MCP MixSearch, you connect your MCP client to the server and issue search requests using the supported tools. You can perform full web searches with content extraction, get concise summaries, or extract content from a specific URL. The server automatically selects the best search engine from multiple backends and applies advanced filtering such as time range, language, geographic region, and content type.

How to install

uv sync

uv run mcp-mixsearch

uv run mcp-mixsearch --mode mcp

uv run mcp-mixsearch --mode rest

Configuration and runtime environment

Configure the server behavior using environment variables. You can place these in a .env file or export them in your shell before starting the server. The key options include host, port, transport mode, authentication, and cross-origin settings. You can also configure Brave’s web search API key and content limits.

# Example environment configuration
MCP_HOST=localhost
MCP_PORT=3000
MCP_TRANSPORT=stdio  # or 'http'
MCP_AUTH_ENABLED=true
MCP_CORS_ORIGINS="http://localhost:3000,http://localhost:5173"

# Web Search Configuration
BRAVE_API_KEY=your_brave_api_key_here
MAX_CONTENT_LENGTH=500000
MAX_CONCURRENT_REQUESTS=5

# Logging
LOG_LEVEL=INFO

Advanced usage and endpoints

The server exposes both an MCP protocol endpoint and, in REST mode, a set of HTTP endpoints with identical functionality. In HTTP mode, you can access the MCP endpoint at /mcp, the documentation at /docs, and health at /health when REST is enabled.

Available tools

full_web_search

Performs a comprehensive web search across multiple engines with content extraction and advanced filtering options.

get_web_search_summaries

Returns lightweight search results with summaries and the same filtering capabilities as full searches.

get_single_web_page_content

Extracts readable content from a specified URL, up to a defined content length.