Page Context & Retrieval

When you open Evelant on a page, the extension scrapes the readable content — targeting the main article or document, stripping out navigation, ads, and other page chrome. It supports both a quick "smart" scan and a "full" scan mode (see Full-Page Deep Scan).

• Chunking — long content is split into chunks so the most relevant snippets can be selected before anything is sent to the model, instead of dumping the entire page in every request.
• Single-page app support — navigation is detected even without a full page reload, so context refreshes correctly as you move around modern web apps.
• Page analysis — for very long pages, content can be compressed server-side into a structured summary before being used as context, keeping requests fast and within token budgets.