Can AI Systems Retrieve JavaScript-Rendered Content Reliably?

Question

If identical content exists across different rendering approaches, does delivery architecture change whether AI systems retrieve, understand, or cite it?

Hypothesis

AI systems will retrieve and cite static HTML more reliably than client-side rendered content.

Expected reliability:

1. Static HTML
2. Server-rendered HTML
3. Client-side rendered JavaScript content

Method

Create three test pages with identical visible content:

• Static Astro page
• Server-rendered page
• Client-side rendered page

Each page should use the same:

• Title
• Headings
• Body copy
• Metadata
• Internal links
• URL depth
• Topic focus

The only variable should be how the content is delivered.

Test Variants

Variant	Rendering Method	Expected AI Accessibility
Static	HTML available immediately	High
Server-rendered	HTML generated before response	High
Client-rendered	Content appears after JavaScript runs	Lower

Evaluation Criteria

Each page will be tested for:

• Retrieval
• Summary accuracy
• Citation behavior
• Content completeness
• Entity recognition

Test Prompts

Example prompts:

• Summarize the test page.
• What is the main recommendation on this page?
• What rendering method does this page use?
• Does this page explain AI retrieval reliability?

Scorecard

Variant	Retrieved	Correct Summary	Cited	Notes
Static HTML	—	—	—	—
Pre-rendered HTML	—	—	—	—
Client-rendered JS	—	—	—	—

Expected Outcome

Static HTML should be easier for AI systems to retrieve, summarize, and cite because the primary content is available immediately in the document.

Client-rendered content may be more fragile because some crawlers and retrieval systems may not execute JavaScript consistently.

Next Step

Build the three test pages and compare how AI systems retrieve and describe each one.

Test Pages

• Static HTML test page
• Pre-rendered HTML test page
• Client-rendered JavaScript test page