Notable engine and product changes, newest first. Silence looks abandoned — so we log every meaningful shipment here.
Every audit now generates a frozen, base64-encoded snapshot at /report — send the URL to your dev team and it renders read-only, with the same fetch receipt. The Evidence Panel below every result now exposes HTTP status, response time, response headers, byte counts, the raw title/description/canonical/JSON-LD snippets we parsed, plus copy-paste curl commands so anyone can reproduce the fetch.
The scoring engine now runs a real server-side fetch of your homepage, /llms.txt, and /robots.txt. Regex-based parsing extracts JSON-LD graph types, meta tags, headings, and content density. Same input, same score — deterministic and reproducible.
The dashboard header now exposes a Re-run button that visibly re-fetches your URL and updates the fetch timestamp. One-click links open Google Rich Results Test and the Schema.org Validator, pre-filled with your URL, so you can cross-check our JSON-LD parsing against the authoritative validators.
The dashboard now leads with the score and the three highest-severity findings. Deep technical modules (Model Matrix, Injection Pipeline, Boardroom Report, Billing) move into an accordion below.
New Generative Engine Optimization tab reports citation footprint and trust signals across ChatGPT, Perplexity, Claude, and Gemini surface areas.
Executive Summary vs Engineering Core toggle at the top of the dashboard — the same audit, two lenses.
Pricing cards no longer stack under each other on narrow viewports.