A Pi extension and standalone CLI for source discovery through SearXNG. It returns normalized, deduplicated search results.
pi install https://github.com/amixaam/web-search.gitThe extension registers one tool: web_search.
git clone https://github.com/amixaam/web-search.git
cd web-search
npm install
npm linkIt uses https://search.amixam.net by default. Override it when needed:
SEARXNG_URL="https://your-searxng-instance" web-search '{"query":"SearXNG"}'web-search '{"query":"latest React release notes", "quantity":5}'
web-search '{"queries":["browser automation", "web scraping"], "category":"news", "time_range":"month"}'
web-search '{"query":"Responses API", "domains":["platform.openai.com"], "language":"en"}'| Parameter | Default | Description |
|---|---|---|
query / queries |
— | A query string, or an array for parallel searches. |
quantity |
3 |
Results per query, from 1–20. |
page |
1 |
First SearXNG result page to inspect. |
category |
general |
Any category enabled by the SearXNG instance, such as images, news, videos, or files. |
time_range |
— | day, week, month, or year. |
language |
all |
SearXNG language code. |
domains |
— | Array of allowed domains; subdomains match. |
The tool strips common tracking parameters, deduplicates URLs, and preserves available engine, publication-date, thumbnail, and image-source metadata.