How is SearXNG different from a VPN or private browser mode?

A VPN hides your IP but your search queries still reach Google's servers. Browser private mode prevents local history but sends every query to your default engine. SearXNG intercepts queries on your server and distributes them across engines — Google never sees your real identity or search pattern.

Can SearXNG power AI agent web search?

Yes. Enable `json` in the `formats` list of `settings.yml`, then query `https://your-domain.com/search?q=query&format=json`. Open WebUI, n8n, Flowise and LangChain all support SearXNG as a web search backend natively.

How much RAM does SearXNG need?

The SearXNG container uses under 256 MB RAM. Redis adds ~30 MB. A 1 GB VPS is sufficient for personal or small-team use; a 2 GB VPS handles multiple concurrent users and heavy AI agent traffic comfortably.

Can I restrict access to my SearXNG instance?

Yes. The simplest method is HTTP Basic Auth configured in your Nginx reverse proxy — only users with credentials can reach the interface. For API-only access, disable the HTML interface in `settings.yml` and restrict IP ranges in Nginx.

Kick off a software project quickly.

SearXNG

Self-hosted private search — aggregate results from 280+ engines with zero tracking. Docker + Redis, under 512 MB RAM, JSON API ready for AI agents.

1 GB (512 MB recommended minimum) RAM 1 vCPU Port 8080 Available

Tech stack

DockerPythonRedis

Minimum RAM1 GB (512 MB recommended minimum)

Minimum CPU1 vCPU

Default port8080

Compatible OSChoice of Linux distributions

SearXNG (AGPL-3.0, 33 k+ GitHub stars, rolling releases) is a free, open-source metasearch engine that proxies your queries to up to 280 upstream search engines — Google, Bing, DuckDuckGo, Brave, Wikipedia, Reddit, Qwant and hundreds more — then merges, de-duplicates and returns the results without tracking, profiling or storing your search history.

Deployed on a ServOrbit VPS, SearXNG becomes your team's private search endpoint: a single URL your developers, AI agents and internal tools query instead of expensive Search APIs — no per-query cost, no rate limits, no data sent to third parties.

The built-in web interface lets users switch between search categories (web, images, news, social, science, files) and engine subsets on the fly. A clean JSON API (`?format=json`) makes SearXNG the ideal search backend for Open WebUI, n8n automations, Flowise agents or any application that needs fresh web results without a paid Search API subscription.

Configuration is purely environment-variable-driven or file-based (YAML): add or remove engines, adjust weights, enable safe-search, configure result proxying, and restrict access — all without restarting the container. Redis is included for rate-limiting and shared result caching.

Key features

Queries 280+ engines simultaneously — Google, Bing, DuckDuckGo, Brave, Wikipedia, Reddit, GitHub, arXiv and more

Zero tracking, no user profiling, no persistent cookies — privacy by architecture, not by policy

JSON API for AI agents — power Open WebUI, n8n, Flowise or any LLM tool with real-time web search

Per-category engine sets — web, images, news, social, science, videos, files; configurable per query

Redis caching — rate-limiting and shared result cache reduce upstream load and improve response times

AGPL-3.0, 33 k+ GitHub stars, weekly Docker image updates — active community, Hacker News-featured in 2026

When to use this solution?

1

AI agent backend

Connect Open WebUI, n8n or Flowise to your SearXNG endpoint and give your AI agents access to real-time web search without API keys or usage quotas. SearXNG's JSON API is natively supported by all major LLM orchestration frameworks.

2

Developer private search

Replace your Google Custom Search API subscription with a self-hosted SearXNG instance. No per-query billing, no rate limits, full control over which engines are queried — and your searches never leave your infrastructure.

3

Team search endpoint

Deploy a shared search instance for your team — every developer and internal tool queries the same endpoint. Results are cached in Redis, your search history stays on your server, and access control is handled by your existing reverse proxy.

Deploy SearXNG on your VPS

Guide optimized for ServOrbit Cloud VPS.

01

Create the VPS

Order a ServOrbit Cloud VPS (minimum 1 GB RAM — SearXNG uses under 256 MB, Redis adds ~30 MB). Choose Ubuntu 22.04.

02

Deploy SearXNG

The provisioner starts two containers: the SearXNG engine and a Redis cache. The web interface is reachable on port 8080 immediately — no initial setup wizard required.

03

Configure Nginx + SSL

Set up an Nginx reverse proxy with Let's Encrypt to expose SearXNG on your domain over HTTPS. Update SEARXNG_BASE_URL to your domain so links in results are correctly generated.

04

Enable the JSON API

Edit /etc/searxng/settings.yml on your VPS and add json to the formats list under search:. Restart the container. Your AI agents can now query https://your-domain.com/search?q=query&format=json.

05

Connect your AI tools

In Open WebUI → Settings → Web Search, set the provider to SearXNG and the URL to https://your-domain.com. In n8n, use the HTTP Request node pointing to the JSON API. Your agents now have live, uncensored web search.

Frequently asked questions

SearXNG is a self-hosted metasearch engine that forwards your queries to up to 280 search engines (Google, Bing, DuckDuckGo, Wikipedia, Reddit, Brave…) and merges the results — without tracking you, storing your history or selling your data.