linkup-search

Choose depth based on where the data lives and whether steps must chain. Use fast for one focused fact, standard for snippet-level answers or scraping one known URL, and deep when you need to scrape multiple URLs or chain search into scraping. When uncertain, default to deep.

Write instruction-style queries that state what to find, where to look, and what to extract, such as asking for the pricing page and specific plan fields. Add context like dates, locations, and domains, and use keyword-style queries only for simple single-fact lookups.

Use /fetch when you already know the exact URL and want its content as markdown, such as scraping a known pricing page or article. Use /search when you do not know which URL holds the answer or need agentic retrieval across multiple pages.

Yes, the /fetch endpoint supports a renderJs option that executes JavaScript before extraction. The Skill recommends defaulting renderJs to true because many sites load content via JavaScript, and the latency tradeoff is worth the reliability gain.

Fast and standard calls cost €0.005 each, while deep calls cost €0.05. For broad research, running 3-5 parallel standard calls with focused sub-queries is often faster and cheaper than a single deep call.

Yes, with a known LinkedIn URL you can extract profile details, recent posts, or comments using a standard-depth call. If you need to discover LinkedIn URLs first and then scrape them, use deep depth with explicit sequential instructions.