SearXNG on a VPS (2026): Private Metasearch Engine

SearXNG is a privacy-friendly metasearch engine you can run yourself.

Table of Contents

1. Why SearXNG
2. VPS specs
3. Install overview
4. Instance settings
5. Maintenance
6. Hardening tips
7. Related guides

1. Why SearXNG

Fast, private search for your team or community.

2. VPS specs

Modest CPU and RAM; SSD for speed.

3. Install overview

Use Docker; configure engines and HTTPS. Create an .env for secrets, map a persistent volume, and place SearXNG behind Nginx Proxy Manager or Caddy for TLS. Keep logs modest and set appropriate rate limits.

4. Instance settings

Tune rate limits and enable privacy options. Disable tracking engines you do not need, reduce outgoing.requests concurrency on small VPS plans, and enable safe search where appropriate for communities.

5. Maintenance

Update regularly; monitor usage. Track CPU and RAM footprints, rotate logs, and back up configuration. Consider Prometheus or Uptime Kuma for basic availability checks.

6. Hardening tips

Run behind a reverse proxy with TLS, enforce strong admin credentials, and set sensible request rate limits. Keep the instance private for small teams, or publish public instances with clear usage policies.

7. Related guides

• Nginx Proxy Manager on a VPS
• Uptime Kuma on a VPS
• Vaultwarden on a VPS
• Immich Photo Server
• See hosting options at /vps-hosting

See /vps-hosting for plans.

Why self-host SearXNG

SearXNG is a metasearch engine: it queries Google, Bing, DuckDuckGo, Brave, Wikipedia, etc. and aggregates results. Your IP never reaches the upstream search engines; they only see your VPS.

What you get:

• No tracking, no profile building.
• No filter bubble (results not personalized to your history).
• Custom search engine for browsers, replacing Google as default.
• API access for scripts and Discord/Slack bots.

What you don't get:

• Better ranking than Google. Aggregation is convenience, not improvement.
• Total anonymity. Your VPS IP is the new identifier; rotate or share.

VPS sizing

SearXNG itself is light. The bottleneck at scale is upstream rate-limits (Google starts CAPTCHAing a single IP after 50-100 queries/min).

docker-compose

services:
  searxng:
    image: searxng/searxng:latest
    container_name: searxng
    ports:
      - "127.0.0.1:8080:8080"
    volumes:
      - ./searxng:/etc/searxng
    environment:
      - SEARXNG_BASE_URL=https://search.example.com/
      - INSTANCE_NAME=my-searxng
    restart: always

  redis:
    image: redis:7-alpine
    restart: always
    command: redis-server --save "" --appendonly no

Reverse proxy 8080 with Caddy:

search.example.com {
  reverse_proxy 127.0.0.1:8080
}

settings.yml essentials

server:
  secret_key: "GENERATE_64_HEX_CHARS"
  limiter: true
  image_proxy: true

search:
  safe_search: 0
  autocomplete: "duckduckgo"
  default_lang: "en"

ui:
  default_theme: simple
  infinite_scroll: false

limiter: true activates the bot detection / rate limiter. Without it a public instance gets scraped within a day.

Public vs private instance decision

Public instances need:

• TLS (Let's Encrypt via Caddy).
• Limiter with Redis.
• Anti-bot rules in nginx/Caddy (rate-limit by IP).
• A privacy policy.

Common breakage

Browser integration

Firefox and Chromium auto-detect SearXNG via OpenSearch. Once visited, right-click the URL bar > Add Search Engine. Set as default for distraction-free, ad-free, profile-free search.