desearch-web-search
par okradze
Web & Frontend DevelopmentRésumé
Search the web and get real-time SERP-style results with titles, URLs, and snippets.
Description
Ce que fait ce skill
Runs real-time web searches via the Desearch API and returns structured SERP-style results with titles, URLs, and snippets. Supports pagination to retrieve additional result pages beyond the first ten.
Returns live web results with structured metadata, unlike static knowledge retrieval that can't reflect recent content.
Quand l'utiliser
- Looking up current news on a topic
- Finding official documentation for a library
- Researching travel destinations
- Checking recent prices or product availability
- Gathering links for a research summary
Exemple de workflow
User asks: find recent articles about large language model benchmarks
- 1
Call desearch.py web with query "large language model benchmarks"
- 2
Receive up to 10 results with titles, snippets, and links
- 3
If more results needed, call again with --start 10 to fetch next page
- 4
Parse returned JSON to extract relevant URLs and snippets
A list of current web results with titles, URLs, and brief descriptions covering recent LLM benchmark coverage
Prérequis
- DESEARCH_API_KEY environment variable
- Desearch account at console.desearch.ai
Installation
clawhub install okradze/desearch-web-searchExécute la commande CLI, ou colle l'instruction dans Claude Code, Cursor, Codex ou l'agent de code que tu utilises.