We relaunched XIA Crawl + Scraper: the web, read and searched by your agents

Context

Scanning a bid document, comparing supplier catalogs, or tracking competitor pricing usually means hours of open tabs, copy-pasting text, and building spreadsheets by hand. Add PDFs dozens of pages long or sites that block access, and the task gets even slower.

This week we relaunched XIA Crawl, the web search and reading module built into NapsixB2B's agentic system. Nothing to install, no own APIs to maintain, no fragile connectors.

What changes

XIA now handles the whole web research cycle in a single flow:

• Native web search: the agent searches and returns up to ten results with title, URL and summary, in seconds.
• Page and PDF reading: scrape_url returns clean Markdown; PDFs — bid documents, catalogs, spec sheets — are read at no extra cost.
• Blocked-site detection: if a site asks for a login or blocks access, XIA reports it instead of serving an error page as if it were valid content.
• Page cache: a page already read comes back in milliseconds without using quota.
• Public API on the Gateway: the same POST /v1/search and POST /v1/scrape endpoints are available with your nsx_live_* key, so you can call them directly from your own code.

What it means

In practice, an 84-page bid document turns into a table with bid guarantee, delivery deadline and required track record. A supplier scan across a sector gets done by reading catalogs and spec sheets from several plants at once. And researching a client ahead of a meeting — history, products, team — takes just one conversation turn.

Every XIA plan includes a daily quota of searches and reads that resets each day, from 20/20 on the Free tier up to unlimited on Enterprise. The Gateway API has its own quota, audited in your usage history.

What we're doing

The module is already live across all XIA plans. Full sites, sitemaps and RSS feeds you index also feed your team's Knowledge Base, available to every agent.

Try it free · See plans · Gateway docs