<b>The render-parity gate: what the bot sees vs. what users see</b>

<b>The render-parity gate: what the bot sees vs. what users see</b>

At scale, a JS-rendering gap means thousands of pages look full to users and empty to crawlers. Make parity a hard gate, sampled continuously.

The parity routine:
☐ Step 1 — Dual fetch. Pull each sampled URL as raw HTML and as fully-rendered DOM. Gate: the primary entity data, H1, and internal links must exist in raw HTML.
☐ Step 2 — Token diff. Compare visible-text token counts. Gate: fail if rendered-only content exceeds 25% of the page — that content is at risk of never being indexed.
☐ Step 3 — Link parity. Every internal link must be a real <code>&lt;a href&gt;</code> in source, not a JS click handler. Gate: fail any link the raw HTML lacks.
☐ Step 4 — Schema parity. JSON-LD present in initial response, not injected post-load. Gate: fail late-injected schema.
☐ Step 5 — Status honesty. Soft-404 detection: an "empty" template that still returns 200 fails.

Guardrail: run parity on 30 random URLs nightly, not once at launch — a frontend deploy can silently break it across the whole set.

Ship gate: don't publish until all boxes are checked.