FREE CHECK
Can AI read your website and recognize your brand?
Enter your website and brand name for a technical check and AI model probes in about a minute. Find the gaps before choosing what to improve. No signup required.
Check Your Brand
SCOPE
What We Check
Ten technical checks total 100 points: a pass earns the full weight, review earns half, and a fail earns zero. Available model probes assess brand recognition separately.
- AI Crawler Access · Weight 20
- Check whether robots.txt allows search and answer crawlers to access this URL.
- Text in Raw HTML · Weight 15
- Text available without running JavaScript is easier for crawlers to retrieve.
- Structured Data · Weight 15
- JSON-LD helps machines identify your organization and the content or services on the page.
- Title and Description · Weight 10
- A clear title and description explain what this page is about.
- HTTPS · Weight 10
- An encrypted connection is a basic requirement for a public website.
- Canonical URL and Language · Weight 8
- A canonical link and html lang identify the primary URL and content language.
- Open Graph Sharing Data · Weight 7
- OG titles, descriptions and images keep page information consistent when shared.
- Heading Structure · Weight 5
- One H1 and clear H2 sections express the page's information structure.
- llms.txt · Weight 5
- This is an emerging content navigation convention. Google Search ignores it, and its absence does not prevent AI discovery.
- Sitemap · Weight 5
- An XML sitemap helps crawlers discover public URLs on your website.
- Do AI Models Recognize Your Brand?
- Two randomly assigned free models say whether they know the brand. When you add a category, we also check whether you appear in their Taiwan recommendations.
METHOD & LIMITATIONS
How to Read This Score
- Each technical check has a fixed rule and weight. Results show measured characters, fields and crawler restrictions. Suggested fixes are ordered by points lost.
- When model results are available, technical checks contribute 60% and models 40%. Without a category, the model score measures recognition only. With a category, recognition and recommendation position each contribute half.
- This is a raw HTML snapshot of one page. We do not run JavaScript or verify that every crawler has visited. Missing or unreadable robots.txt is treated as no known restriction; a firewall may still block crawlers.
- Model answers are not search rankings, exposure measurements or proof of factual accuracy. A high score does not guarantee recommendations. Results with different categories, or without a category, are not directly comparable.
- Websites and models can fail temporarily. Results reflect the check time only. For ongoing measurement, keep prompts, timing and models consistent and repeat the measurements.
Frequently Asked Questions
Does a high score mean ChatGPT will recommend us?
No. The score reflects technical signals on one page and the model answers at that moment. Recommendations also depend on prompts, brand information, competitors, model versions and live search results.
Why can a technically strong site have poor brand recognition?
Readable content does not mean a model has learned about or cited your brand. These model APIs have no live search enabled, so recent information may be missing from their existing knowledge.
Do I have to provide a category?
No. Leaving it blank still runs all technical checks and brand recognition probes. Adding a category asks for five recommended brands in Taiwan and measures whether your brand appears and at what position.
Is blocking GPTBot always a problem?
No. Restricting training crawlers is a legitimate content policy choice, so it receives a review status. Blocking search or answer crawlers in robots.txt receives a fail. Other access controls may still affect actual crawling.
Does missing llms.txt affect Google Search?
Google Search does not use llms.txt. It is an emerging convention worth only five points here and never receives a fail. Prioritize readable content, crawler access and accurate page information.
How does this differ from a full audit?
The free tool checks one page and a small set of model answers. A full AI answer accuracy audit examines broader question scenarios, brand descriptions and cited sources to identify useful content and workflow changes. Tenten engineers and marketers can embed with your team to get a priority improvement live.