SearXNG Hosting

How SearXNG works

SearXNG sends a query to enabled external search engines through adapters, gathers their responses, and combines the result sets into one page. Users can select categories, language, safe-search settings, time ranges, and other options supported by the configured engines. OpenSearch support lets compatible browsers add the instance as a search provider.

The instance does not crawl and maintain a complete independent web index. Results, snippets, images, availability, ordering, latency, and geographic coverage depend on upstream providers and the selected adapters. Providers can change interfaces, apply rate limits, block requests, or return different material, so a self-hosted metasearch page cannot promise fixed source coverage or identical results over time.

Key SearXNG features

Upstream SearXNG is designed to avoid storing search data and to mix queries sent to other platforms. That reduces direct exposure of an individual user to each provider, but the instance operator still controls logs and configuration, while upstream engines see requests from the instance address. Proxy or Tor routing is optional configuration and is not included by the catalogue template.

The catalogue persists /etc/searxng and sets the public base URL. It does not configure a user account system, private access gateway, paid search APIs, proxy network, content moderation policy, or fixed API output formats. Operators should restrict an instance intended for a private audience and review logging, enabled engines, safe-search policy, request limits, and public abuse exposure.

SearXNG vs Google Search

SearXNG is a metasearch interface that aggregates configured providers and avoids creating its own user profile. Google Search runs a provider-managed search service with its own ranking systems and account-linked features. Google documents that personalised recommendations can use account activity, while results may also vary with location, language, device, and the current query.

Google Search may suit people who want its integrated index, specialised result formats, account history, local relevance, and other search services without operating an instance. SearXNG is useful when the priority is to broker queries through a configurable self-hosted front end and accept the reliability and coverage limits of external engines.

Who uses SearXNG

Individuals can make SearXNG their browser default, schools can demonstrate metasearch behaviour, and research teams can compare results across enabled providers from one interface. Organisations can also use a private instance to apply a consistent engine list and presentation for a defined group.

SearXNG is not an independent crawler, contractual search API, legal-discovery system, content archive, anonymity guarantee, or replacement for checking primary sources. Sensitive searches require careful operator logging policy, access control, network configuration, and consideration of what external providers can infer from the instance.

Self-hosting SearXNG: requirements and cost

SearXNG workload depends on simultaneous queries, number and latency of enabled engines, result categories, image proxies, timeouts, rate limits, and public abuse. PostgreSQL and MariaDB are Not required in the catalogue; configuration is persisted under /etc/searxng. The software has no separate application fee, while a widely shared public instance can create much heavier outbound traffic than a private one.

On AvaHost, SearXNG uses Plan 1 at €2. The hosted SearXNG package includes one-click deployment, a custom domain with automated HTTPS, automatic application updates, and scheduled backups. The template sets the base URL and persists configuration but does not add user authentication, proxy or Tor routing, paid provider credentials, moderation, an independent index, or contractual result availability. Upstream engine failures, blocks, latency, and policy changes remain outside the hosting service’s control.