Detect
Compare fetched HTML and rendered DOM to find missing sections.
Guide
How to Read JavaScript Rendered Pages
If static fetch misses key content, switch to browser-rendered extraction.
This guide gives a repeatable process for dynamic pages, including waits, selectors, and fallback checks.
Workflow
Render
Open the page in browser mode and wait for content-bearing selectors.
Extract
Collect structured fields, then verify completeness with spot checks.
Best practices
Prefer semantic selectors
Target stable section headings, labels, and landmarks.
Handle infinite scroll
Paginate or scroll in bounded loops and collect content incrementally.
Handle login walls early
Identify authentication requirements first, then use approved logged-in sessions.
Store extraction schema
Use a fixed output schema to simplify downstream QA.
Related guides
Browser Automation for AI Agents
See where dynamic extraction fits in an agent workflow.
Codex Browser Automation
Apply the same pattern in Codex workflows.