About
A no-frills library to wire real-time web search and scraping straight into your agent-powered apps. TypeScript, REST endpoints, and (soon!) a Python SDK—drop it in and start querying pages or extracting lists of links/images in seconds.
Why it exists
Agent Toolkit wraps search APIs and HTML parsing into one tiny, battle-tested module.
When agents are doing research, they often need to crawl several pages, current tools provided search, but the extraction was missing easy to parse information for agents to browse while doing research. We made this very easy in the extract api by doing a deep extract and examining links in pages (even those without server-side rendering)
Under the hood
Web search via your provider of choice
Result summarization so agents skip the fluff
Page extraction down to link & image lists
Deep scraping hooks for structured data
Give it a spin and let us know what breaks (or what rules we’re missing).