# RobotCheck.coffee — Is your website AI-readable?

> Free analyzer that checks how well search bots, AI crawlers and LLMs can read a website. Returns a score from 1 to 10 with a detailed, actionable report. No signup required.

This is the Markdown version of the homepage https://robotcheck.coffee — the same content, without layout, for LLMs and agents. Site overview for LLMs: [llms.txt](https://robotcheck.coffee/llms.txt).

## Check a website

Enter a domain at [robotcheck.coffee/check](https://robotcheck.coffee/check). The analysis runs instantly and is free (5 checks per day per IP address; more via [API key](https://robotcheck.coffee/pricing)).

## What is checked

The score combines ten weighted criteria:

| Criterion | Weight | What it looks at |
|---|---|---|
| Schema.org markup | 20 % | Structured data (JSON-LD, Microdata) |
| robots.txt | 15 % | Present, well-formed, rules for AI crawlers (ClaudeBot, GPTBot, Google-Extended …) |
| llms.txt | 15 % | Machine-readable site description for LLMs, checked against the current llms.txt rules |
| Sitemap XML | 8 % | Valid sitemap with the important pages |
| Open Graph / Twitter Cards | 8 % | Metadata for link previews |
| Semantic HTML | 8 % | Heading hierarchy and landmark elements |
| Image alt texts | 8 % | Descriptions for images |
| Load time | 8 % | Response time as a proxy for Core Web Vitals |
| Canonical tags | 5 % | Duplicate-content prevention |
| Hreflang | 5 % | Language versions of multilingual sites |

The rules for robots.txt, llms.txt and related HTTP/HTML signals are taken from primary sources (Anthropic, OpenAI, Google, llmstxt.org) and updated weekly. Every report states the rule version it was checked against. Some technical recommendations (e.g. HTTP/2, compression, caching headers) are shown as hints without affecting the score.

## Grades

| Score | Grade |
|---|---|
| 9.0 – 10 | A+ |
| 8.0 – 8.9 | A |
| 7.0 – 7.9 | B+ |
| 6.0 – 6.9 | B |
| 5.0 – 5.9 | C |
| 4.0 – 4.9 | D |
| below 4.0 | F |

## Why machine readability matters

AI systems like ChatGPT, Claude, Gemini and Perplexity read and summarize websites directly. A site they cannot parse well is invisible to this new kind of search — regardless of how good its content is.

## More

- [Rule changelog](https://robotcheck.coffee/en/changelog) — every rule version with sources
- [Why RobotCheck updates its rules automatically](https://robotcheck.coffee/en/blog/why-robotcheck-updates-its-rules-automatically)
- [Blog](https://robotcheck.coffee/en/blog)
- [API access and pricing](https://robotcheck.coffee/en/pricing)
- [Imprint](https://robotcheck.coffee/en/imprint) · [Privacy](https://robotcheck.coffee/en/privacy)