Top fixes
- Readable without JavaScript (+2 points): Server-render or prerender the first paint so the HTML carries your text; agents and most crawlers do not execute your scripts. How to fix
- MCP census entry (+2 points): Expose a Streamable HTTP MCP endpoint at /mcp (or a paid x402 route) so agents can call you, not just read you. How to fix
- Structured data (+1 point): add a JSON-LD block (Organization or WebSite, plus Product, FAQPage or Article where relevant) so answer engines describe you accurately. How to fix
Ordered by points recoverable. Each area: 2/2 pass, 1/2 warn, 0/2 fail.
chat.curie.app is graded D. The $49 AI Visibility Full Report turns the failing check above (structured data) into a prioritized, specific fix list for chat.curie.app, and tests 8 real buyer questions across ChatGPT-class models to show what AI answers say about you today.
Get the full report — $49 Stripe checkout asks for your brand name and category; the report is ready about 60 seconds after payment. Legal notices and the MCP census entry are not in this report (free: Compliance Desk check, census claim).
Crawler access (robots.txt)
2/2 · pass
8 of 8 AI crawlers may read the site, 0 disallowed
- GET /robots.txt -> HTTP 200, 710 B, 10 user-agent groups
- Allowed: GPTBot, ClaudeBot, OAI-SearchBot, PerplexityBot, Google-Extended, Amazonbot (via *) (3 path rules), CCBot (via *) (3 path rules), Bytespider (via *) (3 path rules)
- Sitemap declared: https://shop.curie.io/sitemap.xml
Fix: Nothing to fix here.
Machine-readable surfaces
2/2 · pass
2 of 7 surfaces present (/llms.txt, /sitemap.xml)
- /llms.txt: HTTP 200, 4.4 KB of text
- /sitemap.xml: HTTP 200, XML sitemap (20 KB)
- /.well-known/x402: HTTP 404
- /.well-known/agent-card.json: HTTP 404
- /.well-known/mcp-registry-auth: HTTP 404
- /openapi.json: HTTP 200 but an HTML page (catch-all route, not a real file)
- /ai.txt: HTTP 200 but an HTML page (catch-all route, not a real file)
Fix: Publish one agent surface: /openapi.json, /.well-known/agent-card.json, or /.well-known/x402 if you sell to agents.
Readable without JavaScript
0/2 · fail
Home page is a JS-only shell
- GET https://chat.curie.app/ -> HTTP 200, text/html, 21 KB
- 64 characters of visible text with scripts, styles and tags removed; <title>: "Curie Vision — AI-Powered Shopping | Curie - 3D Product Commerce"
- 29 external scripts, 9 inline; <noscript> absent
- first words: "Curie Vision — AI-Powered Shopping | Curie - 3D Product Commerce"
Fix: Server-render or prerender the first paint so the HTML carries your text; agents and most crawlers do not execute your scripts.
Structured data
1/2 · warn
No JSON-LD, Open Graph complete
- 0 JSON-LD blocks
- Open Graph: og:title "Curie Vision — AI-Powered Shopping"; og:description present
- meta description: present
Fix: add a JSON-LD block (Organization or WebSite, plus Product, FAQPage or Article where relevant) so answer engines describe you accurately.
Legal notices (TAKE IT DOWN desk check)
not measured
Unavailable
- takeitdown.agentexchange.work/check.json: refused takeitdown.agentexchange.work: hosts on agentexchange.work cannot be graded from inside the same Cloudflare zone; every check would - unavailable
Fix: Re-run in a few minutes; the notice check could not be completed.
MCP census entry (index.agentexchange.work)
0/2 · fail
Not selling to agents: no census entry, no /mcp endpoint
- census unavailable (refused index.agentexchange.work: hosts on agentexchange.work cannot be graded from inside the same Cloudflare zone; every check would time )
- GET /mcp -> HTTP 200 HTML page (catch-all route, not an MCP endpoint)
Fix: Expose a Streamable HTTP MCP endpoint at /mcp (or a paid x402 route) so agents can call you, not just read you.
$49 AI Visibility Full Report
This grade is free and automated. The paid next step is the $49 AI Visibility Full Report (which brands a representative AI model, standing in for ChatGPT-class assistants, names when buyers ask about your category, with a prioritized fix list) from aivisibility.agentexchange.work, run for chat.curie.app. It includes:
- 8 real buyer questions tested across ChatGPT-class models
- Competitor share-of-voice: who AI names, how often, versus you
- Full GEO site audit with prioritized, specific fixes
- Agent-Readiness Score: crawler access, llms.txt, schema, discovery manifest
- Custom 30/60/90-day action plan to get cited by ChatGPT, Perplexity and Google AI Overviews
- Shareable report, generated in about 60 seconds after checkout
Get the Full Report — $49 Stripe checkout asks for your brand name and category; the report is ready about 60 seconds after payment. Legal notices and the MCP census entry are not in this report (free: Compliance Desk check, census claim).
