Integration · Crawl
No API? Your sitemap is enough.
AACSearch reads your sitemap or start pages, fetches each page safely, strips the markup and indexes the text. For documentation, it splits pages by headings so a result links to the exact section.
Two ways to crawl
Sitemap crawl
Every URL in sitemap.xml becomes a document: title, text and URL, refreshed on a schedule.
Documentation crawl
DocSearch-style: start URLs, allowed domains and CSS selectors for heading levels and text, so each section is its own result with an anchor.
Safe by design
Public addresses only
The crawler refuses private and loopback addresses and re-checks every redirect, so it can't be turned against an internal network.
Bounded runs
Page and depth limits keep a crawl from running away on a site with infinite calendars or filters.
What it enables
Cited AI answers
The assistant answers from crawled pages and links to the section it used.
Meaning-based search
Hybrid search finds the right article even when the visitor's words differ from the title.
Specifications
- Sources
sitemap.xml, single URLs, documentation start URLs- Pages per docs crawl
- 50 by default, up to 1,000
- Link depth
- 10 by default
- Documents per sitemap run
- 500 by default, up to 2,000
- Redirects followed
- up to 5, each re-checked
- Schedule
- manual, hourly, daily or weekly
Before you start
- A public sitemap or a list of start URLs
- For documentation: CSS selectors of your page title, headings and body
- Pages that render their text in HTML (client-only apps may need the documents API instead)