Category: Research
Node id: searxng
Standards: web-research, metasearch
Run a free-text query against a SearXNG metasearch instance. SearXNG aggregates Google/Bing/DuckDuckGo/Brave/Qwant/etc. behind one JSON API. Each result includes the underlying engine that produced it.
What it does
- Free-text web search via a SearXNG instance
- Category, language, time-range, and safesearch filters
- Per-result engine attribution (which underlying engine produced this hit)
- Optional bearer-auth for protected/commercial SearXNG deployments
Use cases
- Diverse web search without single-engine bias
- Privacy-respecting search (self-hosted SearXNG)
- Compare ranking across engines
Runtime
Output ports
Settings
Basic
Filters
Advanced
Errors
Runtime notes
- results output is empty when status is not ok — downstream nodes must check status
- SEARXNG_BASE_URL and SEARXNG_USER_AGENT env vars are required or the node fails on first Execute
- SEARXNG_API_KEY (optional) is sent as
Authorization: Bearer for protected instances
- Cache TTL is 30 minutes; pass bypass via your service config if you need fresh results
See Build any workflow for how to put nodes together and Ports, types and data flow for how port types connect.