SearXNG Configuration & Features
Search Engine Categories
SearXNG aggregates results from 70+ engines across categories:
| Category | Engines | Example Results |
|---|---|---|
| General | Google, Bing, DuckDuckGo, Brave, Qwant | Web pages |
| Science | Google Scholar, Semantic Scholar, arXiv, PubMed | Academic papers |
| IT | GitHub, Stack Overflow, Arch Wiki, Hacker News | Code & docs |
| Images | Google Images, Bing Images, Flickr, Unsplash | Photos |
| Videos | YouTube, PeerTube, Dailymotion | Videos |
| News | Google News, Bing News, Reuters, AP | News articles |
| Maps | OpenStreetMap, Photon, Nominatim | Locations |
| Music | SoundCloud, Bandcamp, Radio Browser | Audio |
| Files | 1337x, Piratebay, Nyaa | Torrents |
JSON API for AI Applications
SearXNG provides a clean JSON API — perfect for AI pipelines:
# Search and get JSON results
curl "http://localhost:8080/search?q=AI+agents+2026&format=json"import requests
def search(query, categories="general", num_results=10):
resp = requests.get("http://localhost:8080/search", params={
"q": query,
"format": "json",
"categories": categories,
"pageno": 1,
})
results = resp.json()["results"][:num_results]
return [{"title": r["title"], "url": r["url"], "content": r["content"]}
for r in results]
# Use in your AI pipeline
results = search("best practices for RAG pipelines")
context = "\
".join(f"- {r['title']}: {r['content']}" for r in results)
# Feed context to your LLMIntegration with AI Search Apps
SearXNG is the default search backend for major AI search tools:
# Morphic .env
SEARXNG_API_URL=http://localhost:8080
# Perplexica .env
SEARXNG_URL=http://localhost:8080
# Open WebUI — enable in admin settings
# Settings > Web Search > SearXNG URL: http://localhost:8080Settings Configuration
# searxng/settings.yml
use_default_settings: true
general:
instance_name: "My Search"
search:
safe_search: 0
default_lang: "en"
autocomplete: "google"
server:
secret_key: "your-secret-key-here"
bind_address: "0.0.0.0"
port: 8080
engines:
- name: google
engine: google
shortcut: g
disabled: false
- name: github
engine: github
shortcut: gh
disabled: false
- name: arxiv
engine: arxiv
shortcut: ax
disabled: falsePrivacy Features
- No tracking — Zero cookies, no user profiles, no search history on server
- No ads — Clean results without sponsored content
- Proxy support — Route image and result links through SearXNG
- Tor-friendly — Works as a Tor hidden service
- Self-hosted — Your data never leaves your infrastructure
FAQ
Q: What is SearXNG? A: SearXNG is a free, self-hosted metasearch engine with 27,500+ GitHub stars that aggregates results from 70+ search providers without tracking. It's the most popular search backend for AI applications like Morphic and Perplexica.
Q: How is SearXNG different from Google or DuckDuckGo? A: SearXNG aggregates results from multiple engines (including Google and DDG) but strips all tracking. Being self-hosted, there are no rate limits, no API keys needed, and no costs — you control the entire search infrastructure.
Q: Is SearXNG free? A: Yes, fully open-source under AGPL-3.0. Self-host on your own infrastructure at zero cost.