Provides containerization support for easy deployment of the MCP server with configurable environment variables for connecting to SearXNG instances.
Enables language models to perform web searches through SearXNG, allowing search queries with customizable parameters like categories, page numbers, time ranges, and result formatting.
SearXNG Model Context Protocol Server
A Model Context Protocol (MCP) server for interfacing language models with SearXNG search engine.
Description
This server enables language models to perform web searches through SearXNG using the Model Context Protocol standard. It provides a clean interface for language models to send search queries to SearXNG and receive formatted results.
Related MCP server: OneSearch MCP Server
Installation
Requirements
Node.js 16+
A running SearXNG instance (by default at http://localhost:8888)
Usage
The server will run on stdio, making it suitable for integration with MCP-compatible language models.
Configuration
The server can be configured using environment variables:
Variable | Description | Default |
SEARXNG_PROTOCOL | Protocol to use (http/https) | http |
SEARXNG_HOST | SearXNG host | localhost |
SEARXNG_PORT | SearXNG port | 8888 |
CACHE_TTL | Cache time-to-live in milliseconds | 600000 (10 minutes) |
MAX_CACHE_SIZE | Maximum number of cached queries | 100 |
Example:
Docker
The project includes a Dockerfile for easy deployment:
Search Parameters
The search function supports the following parameters:
query(required): The search query stringcategories: Comma-separated list of search categoriespageno: Search page number (default: 1)time_range: Time range for results ("day", "week", "month", "year")raw_json: Return raw JSON response instead of formatted text (default: false)
License
MIT