🔍 SearXNG for AI Agents: Private Search Infrastructure

⚡ Overview

A privacy-respecting, metasearch engine infrastructure designed specifically for AI Agents and LLMs (Large Language Models).

This setup provides a sanitized, JSON-structured output from multiple search engines (Google, Bing, DDG, etc.) without tracking, rate-limiting (via rotation), or ad clutter, making it the ideal Knowledge Retrieval Tool for RAG (Retrieval-Augmented Generation) pipelines.

🏗 Architecture

The stack is composed of three lightweight services orchestrated via Docker Compose:

• SearXNG: The metasearch core engine.
• Valkey (Redis Fork): High-performance caching and rate-limiting management.
• Caddy: Reverse proxy with automatic HTTPS termination.

🚀 Quick Start

Prerequisites

• Docker & Docker Compose
• Git

Installation

1. Clone the repository:

   git clone [https://github.com/dacatano/SearXNG.git](https://github.com/dacatano/SearXNG.git)
   cd SearXNG

2. Configure Environment: Create a .env file from the example template.

   cp .env.example .env
   # Edit .env to set your hostname and generate a secure secret

3. Deploy:

   docker compose up -d

4. Verify: Access http://localhost:8080.

🤖 AI Agent Integration

To retrieve results in JSON format suitable for LangChain, AutoGPT, or custom agents, append &format=json to the query.

Example Endpoint:

GET http://localhost:8080/search?q=latest+advancements+in+LLMs&format=json

🛠 Configuration
This project follows the 12-Factor App methodology.

Secrets: Injected via .env (not hardcoded) in the docker-compose.yml file.

Persistence: Managed via Docker Named Volumes (valkey-data, searxng-data).

Logs: JSON-driver enabled for seamless integration with observability stacks (ELK/Loki).

For architectural decisions and trade-offs, see docs/ARCHITECTURE.md.