Buyer
I want the outcome. You sold me the architecture.
If I already know this category, I can squint and see value. If I don't, I bounce.
Roast Agent
AI users experience your product, find what's confusing, and tell you what to fix.
$29 per Roast. One time.
Sample report
Roast Score
61
Your hero sounds impressive. I still don't know what you sell.
Clarity
48
Conversion
54
Trust
67
UX
72
Differentiation
63
AI Slop
41
Agentic
34
Your biggest problem
The first screen promises an operating system. A first-time visitor is still shopping for a product.
AI slop
41
The homepage sounds like it was generated by asking ChatGPT for a Series A landing page.
100 is distinctive and human. 0 is an interchangeable GPT landing page.
Agentic
34
Agents can load the page. They cannot really use it.
Essential 3/11 · Recommended 4/21
The homepage HTML is mostly an app shell. An agent that does not execute JS gets almost nothing.
No /llms.txt found. Agents have no machine-readable site guide.
Unknown paths return the app shell with HTTP 200, so agents think every URL exists.
100 means agents can find, read, and use this site. Scored by Is Agentic / Ora. Not part of the Roast Score.
Full agentic reportBuyer
If I already know this category, I can squint and see value. If I don't, I bounce.
Skeptic
Claims outrun proof. Logos would help. A named result would help more.
Distracted User
The primary CTA is visible. The product is not.
1. positioning · critical impact · minutes
The hero is atmospheric instead of specific.
2. navigation · high impact · minutes
Pricing is buried under 'Learn more'.
3. trust · high impact · hours
Trust is implied, not demonstrated.