Render Diff.

The AEO metric that matters in 2026

Google's crawler now renders JavaScript reliably. Most AI search crawlers do not. GPTBot, Claude-Web, PerplexityBot, CCBot all default to a non-rendering fetch — they see your static HTML, never the post-hydration DOM. If your content lives in client-rendered React components, your AEO presence is whatever's in the initial HTML payload — typically a near-empty shell.

render-diff measures the gap. It fetches the page twice in parallel — once with a plain reqwest GET, once through a headless Chrome render — and diffs the results.

curl -X POST https://api.crawlcrawl.com/v1/actors/render-diff \
  -H "Authorization: Bearer crk_..." \
  -d '{"url":"https://your-spa.example.com"}'

# → 200
{
  "actor": "render-diff",
  "elapsed_ms": 4444,
  "data": {
    "static":   { "word_count": 12,   "link_count": 0,   "json_ld_count": 0 },
    "rendered": { "word_count": 2482, "link_count": 39,  "json_ld_count": 2 },
    "delta": {
      "ai_bot_blind_pct": 0.995,
      "assessment": "severe — most content invisible to non-rendering crawlers",
      "links_only_after_render": [ "..." ],
      "headings_only_after_render": [ "..." ]
    },
    "rendered_cost_usd": 0.0024
  }
}

The assessment scale

ai_bot_blind_pct	Assessment	Interpretation
≥ 0.5	severe	Most content invisible to non-rendering crawlers. SPA-only site.
0.2 – 0.5	moderate	Significant content requires JS. Above the fold OK, below the fold lost.
0.05 – 0.2	minor	Small share of content is JS-rendered. Mostly OK for AI.
< 0.05	negligible	Static HTML matches rendered DOM. Full AI visibility.

What's compared

• Links — set difference. Surfaces links injected by client-side routers (Next.js / React Router) that AI crawlers never follow.
• Headings — h1-h6 set diff. Catches the "we built our hero entirely in client-rendered JSX" failure mode.
• JSON-LD blocks — count diff. If schema is only injected post-render, every AI ingestion misses it.
• Visible-text word count — drives the headline metric.
• HTML bytes — for context (often paradoxical: rendered is sometimes smaller after image lazy-loading).

When to use it

AEO migration checklist. Before pushing SPA-style pages, run render-diff against the staging URL. Any page with ai_bot_blind_pct > 0.2 needs SSR/SSG or a fallback strategy.

Competitive AEO. Compare your render-diff to a competitor's. If theirs scores 0.0 and yours scores 0.6, you have a structural gap to AI visibility that no amount of content writing fixes.

Periodic monitoring. Combine with Monitors to track render-diff over time. A regression to "severe" usually means a deploy broke SSR.

Pricing

One page-credit + one cloud-credit per call (covers the chrome render). Typical cost per call: ~$0.0024 USD. See full pricing →

Where it fits