Web Search

Web Search (web-search) is a built-in, platform-managed capability — it's available to Webbee 🐝 in every chat automatically, with nothing to install. When you ask a question that needs current information, Webbee searches the live web and reads the best pages for you, then answers from what it found.

Researching the web

Choosing the right depth

• A quick answer from search results — Webbee starts with web_search; the result snippets may already answer the question.
• One or several articles — read_url is the normal choice for one page; read_urls checks a small set in parallel when comparison matters.
• A page that depends on browser rendering — read_url_rendered reads the rendered page when the normal reader cannot see the content.
• A visual or interaction check — capture_page is for seeing what a person sees, including a screenshot and useful browser diagnostics. It is not used as a substitute for article reading.
• A document linked from the web — read_document handles supported Office documents.

Webbee chooses the lightest method that can answer your question, and follows your saved deep-read policy for the heavier readers.

What it costs

Web research runs like any other action — each call costs its credit price plus the platform fee, shown before it runs (see Billing & Earnings). The two token-heavy readers (read_url_rendered, read_document) cost more than a plain read_url; set_web_read_policy lets you cap how often Webbee reaches for them.