Answer-Engine Audit: four engines, one input
One URL, four verdicts. We read your site the way each engine does - its crawlers, its retrieval, its extraction - and show, per engine, what is missing before it can read, trust, and cite you.
- No signup
- About 10 seconds
- Report + workspace by email
No signup · we read your home page, robots.txt, llms.txt and sitemap only · nothing stored unless you add your email
What the audit actually checks
ChatGPT has two doors into answers. OAI-SearchBot decides what ChatGPT Search may cite; GPTBot decides what it already knows about your category. We read your robots.txt for both, then check whether your facts are extractable - a plain "[Brand] is a …" definition, schema.org it can trust, and an llms.txt invitation.
Gemini grounds its answers in Google Search, so it sees exactly what Googlebot indexes. We check Googlebot access, whether Google-Extended is welcome (that is what lets your pages be used for grounding), whether your entity survives as structured data, and the mobile-first basics.
Google AI Overviews lifts passages into the results themselves. Same crawl
prerequisites as Gemini, plus the two things Overviews leans on: discoverable freshness
(sitemap lastmod, dateModified, article:modified_time) and
answer-shaped sections - question headings, FAQPage schema, passages that read like conclusions.
Microsoft Copilot grounds on Bing. It is the one engine where classic technical still wins: Bingbot access, a discoverable sitemap, well-sized titles and descriptions, and structured data. Teams that never touched Bing Webmaster lose Copilot before it ever answers.
These are readiness signals, not answers - no free flag can tell you what the engines say today. When the doors are open, start a trial and see the actual answers, across all eight default engines.
Frequently asked
Which answer engines does this audit cover?
Four: ChatGPT, Gemini, Google AI Overviews, and Microsoft Copilot - the surfaces most buyers hit first. Each gets its own score built from what that engine specifically needs: its crawlers in robots.txt, its retrieval prerequisites, and the on-page signals its systems extract. The trial scan adds Claude, Perplexity, Grok, and Kimi with verbatim answers.
Is this the same as the trial answer scan?
No, and that is the point. This is a readiness audit: whether each engine CAN read, trust, and cite you - crawler access, freshness, structured facts, answer-shaped content. The trial scan shows what the engines actually SAY: verbatim answers, your position, which competitors they name instead.
Why do you ask for my email?
Optional, and it does one thing: your workspace - this audit, the scores, the fix list - is reserved against it, and the full report goes to your inbox. That's the whole account. No password until you continue into the trial, no newsletter, no card. Leave it empty and the audit runs anyway.
What happens when I continue into the workspace?
The trial account attaches to the workspace this audit is saved in, so the report and the fix list are already there when you land. The link carries a workspace id and your domain; a first-party cookie (7 days) bridges the hop from sonarvue.com to app.sonarvue.com. No retyping, no card.
Do these scores mean the engines recommend me?
Not yet - and beware any free tool that implies otherwise. These scores mean the doors are open: the right crawlers are allowed, your facts are extractable, your content is fresh and answer-shaped. Whether an engine actually names you for a buyer prompt is the answer-level question, and that needs repeated, captured runs.