How do I configure the SearXNG URL?

Set the environment variable SEARXNG_URL (e.g., export SEARXNG_URL="http://localhost:8888"). If unset, the default http://localhost:8888 is used.

What is the default result format and how can I customize it?

The script returns titled results with URLs and content snippets. You can customize by passing filters like category, time_range, language, and num_results in a JSON payload.

What about errors and rate limits?

The script handles common errors (network timeouts, unavailability, invalid queries, rate limiting, and empty results). It enforces a minimum 1-second delay between requests and uses a 30-second timeout; a 429 response indicates rate limiting.

Searxng Web Search

Scanned

@noblepayne

npx machina-cli add skill @noblepayne/searxng-local-search --openclaw

Files (1)

SKILL.md

3.6 KB

SearXNG Web Search

Search the web using a self-hosted SearXNG instance. This skill provides access to web search results through the SearXNG JSON API, with built-in rate limiting, error handling, and result formatting.

When to Use

Use this skill when you need to:

Find current information or recent news
Research topics beyond your knowledge cutoff
Look up documentation or technical references
Verify facts or check current status
Find URLs or resources on specific topics
Search for code examples or solutions

Configuration

Set the SEARXNG_URL environment variable to your SearXNG instance:

export SEARXNG_URL="http://localhost:8888"

Or use the default (http://localhost:8888) if not set.

Usage

Execute the search script with your query:

bb scripts/search.clj "your search query"

Advanced Options

Pass additional parameters as JSON:

bb scripts/search.clj "your query" '{"category": "news", "time_range": "day", "num_results": 10}'

Available options:

category - Filter by category: general, news, images, videos, it, science
time_range - Time filter: day, week, month, year
language - Language code (default: en)
num_results - Number of results to return (default: 5)

Output Format

The script returns formatted search results as text:

Search Results for "your query"
Found 42 total results

1. Result Title [Score: 1.85]
   URL: https://example.com/page
   Description snippet from the page...
   Engines: google, bing

2. Another Result [Score: 1.62]
   ...

Error Handling

The script handles common errors gracefully:

Network timeouts (30s timeout)
SearXNG unavailable (clear error message)
Invalid queries (error details)
Rate limiting (429 responses)
Empty results (informative message)

Rate Limiting

The script implements basic rate limiting:

Minimum 1 second between requests
Uses filesystem-based state (.searxng-last-request)
Prevents accidental spam

Examples

Basic Search

bb scripts/search.clj "NixOS configuration"

News Search

bb scripts/search.clj "AI developments" '{"category": "news", "time_range": "week"}'

Technical Search

bb scripts/search.clj "babashka http client" '{"category": "it", "num_results": 3}'

Recent Results Only

bb scripts/search.clj "product launch" '{"time_range": "day"}'

Troubleshooting

"SEARXNG_URL not set"

Set the environment variable: export SEARXNG_URL="http://localhost:8888"

Connection timeout

Check that SearXNG is running: curl $SEARXNG_URL/search?q=test&format=json
Verify firewall settings
Check service status: systemctl status searx

Empty results

Try a broader query
Remove filters and try again
Check SearXNG logs: journalctl -u searx -n 50

Rate limit errors

Wait a few seconds between searches
The script enforces minimum 1s delay automatically

Implementation Notes

The search script (scripts/search.clj) uses:

babashka.http-client for HTTP requests
Clojure's cheshire.core for JSON parsing
Filesystem-based rate limiting
30-second timeout with proper error messages
Result scoring and sorting for best results first

For detailed API documentation, see references/api-guide.md.

Source

git clone https://clawhub.ai/noblepayne/searxng-local-searchView on GitHub

Overview

SearXNG Web Search queries a self-hosted SearXNG API to return ranked results with titles, URLs, and content snippets. It’s ideal for current information, research topics, documentation lookup, and fact-checking beyond your knowledge cutoff.

How This Skill Works

The skill sends your query to SEARXNG_URL via the JSON API, then handles HTTP requests with babashka.http-client, applies built-in rate limiting, and formats the results with scores, titles, URLs, and snippets for easy scanning.

When to Use It

Find current information or recent news
Research topics beyond your knowledge cutoff
Look up documentation or technical references
Verify facts or check current status
Find URLs or resources on specific topics

Quick Start

Step 1: Set the SearXNG URL: export SEARXNG_URL="http://localhost:8888"
Step 2: Run a query: bb scripts/search.clj "your search query"
Step 3: Optional: pass filters as JSON like: bb scripts/search.clj "your query" '{"category": "news", "time_range": "week"}'

Best Practices

Set SEARXNG_URL correctly in your environment
Use category, time_range, and language filters to narrow results
Tune num_results to balance breadth and speed
Rely on the description snippets to judge relevance
Respect rate limits (minimum 1s between requests, tracked by a local state file)

Example Use Cases

NixOS configuration search (basic)
AI developments — news search with category: news and time_range: week
babashka http client — IT category, num_results: 3
product launch — time_range: day
Look up library documentation (general documentation search)

Frequently Asked Questions

Add this skill to your agents