AI companies: how AI-ready are the leaders?
Model labs and AI-native products — the companies building answer engines. We ran the LLMready crawler against each homepage and scored it the same way as a free audit.
73/100
Median (all industries: 77)
60%
Publish an llms.txt
90%
Allow all 4 core AI crawlers
50%
Use Schema.org JSON-LD
Key findings
- Jasper leads with 86/100 (B); Midjourney trails at 49.
- The ai companies median is 4 points below the median of all 102 sites we measured.
- 1 of 10 block at least one core AI crawler (GPTBot, PerplexityBot, ClaudeBot or Google-Extended) in robots.txt.
- 60% publish an llms.txt and 0% an llms-full.txt. 10% rely on JavaScript to render their main text.
- Weakest category on average: citation readiness at 59/100.
Bot access
82
Content extractability
68
Semantic metadata
76
Citation readiness
59
Scores
Leaderboard
| # | Site | Score | Bot access | Extractability | Metadata | Citation | Core bots | llms.txt | Biggest gap |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 86 Jasperjasper.ai | 86 B | 90 | 79 | 97 | 78 | Authorship and freshness | ||
| 2 | 84 ElevenLabselevenlabs.io | 84 B | 84 | 82 | 97 | 76 | Answer-ready formats | ||
| 3 | 83 Mistral AImistral.ai | 83 B | 94 | 74 | 100 | 66 | Authorship and freshness | ||
| 4 | 82 Runwayrunwayml.com | 82 B- | 94 | 79 | 92 | 64 | Authorship and freshness | ||
| 5 | 78 Coherecohere.com | 78 C+ | 90 | 67 | 87 | 66 | Authorship and freshness | ||
| 6 | 68 Anthropicanthropic.com | 68 D | 76 | 72 | 67 | 55 | llms.txt is present and valid | ||
| 7 | 68 Replicatereplicate.com | 68 D | 90 | 76 | 43 | 53 | Schema.org JSON-LD | ||
| 8 | 66 Hugging Facehuggingface.co | 66 D | 76 | 61 | 61 | 63 | llms.txt is present and valid | ||
| 9 | 52 Perplexityperplexity.ai | 52 F | 76 | 30 | 70 | 31 | Content is in the initial HTML | ||
| 10 | 49 Midjourneymidjourney.com | 49 F | 49 | 61 | 48 | 39 | 2 | llms.txt is present and valid |
Not scored because they turned away our crawler: OpenAI (blocked automated requests (HTTP 403)), Character.AI (blocked automated requests (HTTP 403)).
Most common gaps
Where ai companies sites lose points
Answer-ready formats
100%10 of 10 sites · Citation readiness
Add an FAQ section (with FAQPage schema) and question-style headings that mirror real queries.
Authorship and freshness
100%10 of 10 sites · Citation readiness
Show an author and published/updated dates (in markup and JSON-LD).
Content density
100%10 of 10 sites · Content extractability
Trim inline scripts, styles and markup bloat so text makes up a larger share of the page.
llms-full.txt is available
100%10 of 10 sites · Bot access
Publish the generated llms-full.txt alongside llms.txt.
Markdown readability
90%9 of 10 sites · Content extractability
Shorten sentences and replace jargon; aim for a reading ease of 50+.
llms.txt is present and valid
80%8 of 10 sites · Bot access
Fix the validation issues below or regenerate the file with LLMready.
Where would your site rank?
Run the same audit on your site and compare it to the ai companies median of 73.
A high score means AI engines can reach, read and understand a site. It doesn't guarantee citations — those also depend on authority, relevance and the question being asked.