SearXNG : Core Philosophy, Key Features, Advantages, and How to Use Public & Private Instances
👤 tarun basu •
📅 April 5, 2026 •
👁️ 2 views
• 🔄 Updated April 15, 2026
opensource
selfhosted
## What Is **SearXNG**?
**SearXNG** is a free and open-source **metasearch engine** that aggregates results from many different search services while **not tracking, storing, or profiling users**. It was created as the successor to the original *Searx* project, continuing its philosophy of privacy and decentralization.
| Feature | Details |
| --- | --- |
| Type | Metasearch engine (doesn’t crawl/index by itself) |
| License | GNU Affero General Public License (AGPL-3.0) |
| Codebase | Python (with web interface in HTML/CSS) |
| Runs on | Linux, macOS, FreeBSD, and other Unix-like systems |
| Offered as | Public instances or self-hosted servers |
| Released | March 27, 2021 |
| Website & Docs | [docs.searxng.org](http://docs.searxng.org) and public instance listings at [searx.space](http://searx.space) |
## How It Works
SearXNG **does not function like traditional search engines** that crawl and index web pages itself. Instead:
You enter a search query in the SearXNG interface.
SearXNG sends that query to multiple configured search engines (e.g., Google, Bing, DuckDuckGo, Qwant) simultaneously.
It **aggregates, deduplicates, and cleans up** the results before showing them to you.
At no point does SearXNG store or log personal queries or user IP addresses.
This aggregation means you often get **broader and less biased results** than you would from any single engine.
## Strong Privacy & Anonymity
Privacy is the core philosophy of SearXNG:
| Feature | Description |
| --- | --- |
| No Tracking | SearXNG does not log search queries, IP addresses, cookies, or user data. |
| Opaque to Upstream Engines | Requests to search engines remove private data before being sent. |
| No Profiling | No personalization or ad-targeting based on user activity. |
| Optional Tor Support | Searches can be routed through Tor for additional anonymity. |
| Encrypted Connections (HTTPS) | All traffic is encrypted to ensure security and privacy. |
## Core Philosophy and Privacy
SearXNG’s primary goal is to provide search results without compromising user privacy. Unlike mainstream search engines that track queries, it implements a “zero-tracking” policy.
**No User Tracking**: It does not store or share search queries, and by default, it disables all tracking cookies.
**IP Protection**: All search requests are routed through the SearXNG instance, which hides your real IP address from the external search engines it queries .
**Result Sanitization**: The engine actively removes tracking parameters from the URLs of search results before presenting them to you.
**Forced Encryption**: It strongly encourages the use of HTTPS to encrypt the connection between you and the instance, ensuring your search activity cannot be easily intercepted
## Key Features and Customization
SearXNG offers a highly customizable search experience that puts you in control.
**Powerful Search Syntax**: You can refine your searches with special commands directly in the search box .
Use `!` to search specific engines or categories (e.g., `!wp paris` to search Wikipedia, or `!images car` for an image search).
Use `:` to select a language filter (e.g., `:fr !wp Wau Holland`).
It supports **external bangs** from DuckDuckGo with `!!` to search other sites directly (e.g., `!!wfr Wau Holland`).
**Extensive Filtering Options**: You can filter searches by time (day, week, month, year), language, and safe search level to fine-tune your results.
**Multiple Categories**: Search across different categories like general web, images, videos, news, and maps, often with the ability to select multiple categories at once.
**Customizable Interface**: You can change the theme, including a dark mode, and adjust the layout to your liking. You can even set results to open in a new tab by default.
**Special Queries**: It includes built-in tools for tasks like generating a random UUID, calculating averages, or hashing text (e.g., `md5 lorem ipsum`).
## Technology and Architecture
SearXNG is built on a modular and extensible architecture.
**Metasearch Framework**: It acts as an aggregator, sending your query to multiple search engines simultaneously and merging the results. This is powered by a plugin-like system of **engine adapters** for each external service.
**Performance Focus**: To ensure fast searches, it uses **asynchronous requests** to query engines in parallel. It can also be configured with a caching system like **Redis** to store popular queries and reduce load.
**Flexible API**: It provides a RESTful API, allowing developers to integrate its search capabilities into other applications or tools.
## How to Get Started with SearXNG
You have two main options for using SearXNG: use an existing public instance or run your own.
**Use a Public Instance**: This is the easiest way to try it out. You can find a list of publicly available instances on the SearXNG project website. You can even set one as your browser’s default search engine.
**Run Your Own Instance**: For maximum privacy and control, you can self-host SearXNG. This is surprisingly accessible thanks to its official Docker image.
**Docker Deployment (Recommended)**: The quickest method is using Docker. After installing Docker, you can run a container with a simple command :
bash
docker run --name searxng -d -p 8888:8080 [docker.io/searxng/searxng:latest](http://docker.io/searxng/searxng:latest)
You would then access your new instance at `http://localhost:8888`.
**Other Deployment Options**: You can also install SearXNG directly on a Linux server (like Ubuntu) with or without Docker. For Windows users who prefer not to use a virtual machine or WSL, community projects offer adapted versions that run natively.
## Search Capabilities
### Metasearch Engine Aggregation
SearXNG can query *hundreds* of separate search engines, including:
Standard web search sources
Images, videos, and news aggregators
Specialized sources (scientific, social media, music, files, etc.)
Multiple language and country variations
Some instances support **up to around 247 search engines**, with many enabled by default.
### Search Categories
Users can choose to search within different categories, such as:
General web, Images, Videos, News, Social Media, Files, Science & academia, Maps and more
This categorization helps narrow results to specific content types.
## Customization & Control
### Self-Hosting
You can **run your own SearXNG instance** on a server, local machine, or cloud provider. This gives *full control* over:
Which engines are queried
UI themes and languages
Request handling and privacy settings
Maximum anonymity without relying on public instances
Self-hosting is popular among privacy-focused individuals and organizations.
### User Preferences
Even on public instances, users can customize:
Search categories and engines
Language filters
Themes and appearance
Default search engine settings (via OpenSearch)
## Architecture & UI
Built on **Flask**, a lightweight Python web framework.
UI renders server-side HTML with progressive enhancement so core search works without heavy JavaScript.
Multiple themes, responsive design, and custom icon systems for different result types.
There’s a clear separation between request handling, engine loading, query orchestration, and result formatting.
## Advantages
**Privacy-first search** — no tracking or profiling.
**Aggregated results from many sources** gives broader insights.
**Open Source & Transparent** — code is open for audit and contributions.
**Self-hostable & customizable** for users and organizations.
**Supports Tor** for enhanced anonymity
## Considerations (Limitations)
**Public instance reliability** can vary — they’re run by volunteers.
**Self-hosting requires technical knowledge** (Docker, server setup).
**Advanced search operators & deep query features** may be limited compared to Google/Bing.
Results may sometimes differ in quality or relevance, especially for non-English queries depending on engine configurations. (community feedback)
## Use Cases
Typical use cases include:
**Privacy-first web searching**
**Custom search portals** for companies or communities
**Integration with workflows or AI tools**
**Educational and research purposes**
**Comparative search results across sources**
## Public & Private Instances
**Public instances**: Everyone can use them like a typical search engine.
**Private/self-hosted instances**: Only you or your organization uses them — best for maximum privacy.
A list of public instances is maintained at *searx.space*.
## Community & Development
SearXNG is actively developed by a community of contributors.
Code and issue tracking hosted on GitHub.
There are regular updates, plugins, and community-driven enhancements.