Step 1 of 5 — Run the AVR self-audit via curl. Paste your domain and get back section scores + a verdict in JSON.
# Run the AVR self-audit against your site
curl -s 'https://chudi.dev/api/webmcp/avr-self-audit?format=full&url=yourdomain.com' \
| jq .
# Or via the WebMCP protocol (AI-readable endpoint):
curl -s 'https://chudi.dev/api/webmcp' \
-X POST \
-H 'Content-Type: application/json' \
-d '{"method":"avr_self_audit","params":{"url":"yourdomain.com"}}'Step 2 of 5 — The AVR CLI runs 15 checks across SEO Foundation + AI Infrastructure, scores each section, and prints the Verdict.
Step 3 of 5 — The REST API response: section scores, fact-block density, remediation priorities, and Verdict: AI-READY.
| accept | application/json |
|---|---|
| user-agent | curl/8.7.1 |
| cache-control | no-store |
|---|---|
| content-type | application/json; charset=utf-8 |
| x-avr-verdict | AI-READY |
| x-avr-version | 1.2.0 |
{
"url": "example.com",
"audited_at": "2026-06-05T10:00:00Z",
"sections": {
"seo_foundation": {
"score": 6,
"total": 7,
"pct": 86,
"checks": [
{
"id": "1.1",
"label": "title tag present & unique",
"result": "pass",
"detail": "60 chars"
},
{
"id": "1.2",
"label": "meta description present",
"result": "pass",
"detail": "148 chars"
},
{
"id": "1.3",
"label": "canonical URL declared",
"result": "pass",
"detail": null
},
{
"id": "1.4",
"label": "h1 present on key landing pages",
"result": "warn",
"detail": "2 of 5 pages missing"
},
{
"id": "1.5",
"label": "sitemap.xml accessible",
"result": "pass",
"detail": null
},
{
"id": "1.6",
"label": "robots.txt present & non-blocking",
"result": "pass",
"detail": null
},
{
"id": "1.7",
"label": "https + valid cert",
"result": "pass",
"detail": null
}
]
},
"ai_infrastructure": {
"score": 6,
"total": 8,
"pct": 75,
"checks": [
{
"id": "2.1",
"label": "JSON-LD schema.org markup present",
"result": "pass",
"detail": "Organization, WebSite"
},
{
"id": "2.2",
"label": "llms.txt present at /llms.txt",
"result": "pass",
"detail": "847 bytes"
},
{
"id": "2.3",
"label": "llms-full.txt present",
"result": "warn",
"detail": "file missing"
},
{
"id": "2.4",
"label": "OpenGraph tags complete",
"result": "pass",
"detail": "og:title, og:description, og:image"
},
{
"id": "2.5",
"label": "fact-block density >=20 per page",
"result": "pass",
"detail": "avg 24 extractable facts/page"
},
{
"id": "2.6",
"label": "author entity markup (Person)",
"result": "warn",
"detail": "missing on 3 blog posts"
},
{
"id": "2.7",
"label": "citation-safe prose style",
"result": "pass",
"detail": null
},
{
"id": "2.8",
"label": "WebMCP endpoint accessible",
"result": "pass",
"detail": "HTTP 200 in 87ms"
}
]
}
},
"fact_block_density": {
"extractable": true,
"avg_facts_per_page": 24,
"threshold": 20
},
"summary": {
"total_checks": 15,
"passed": 12,
"warned": 3,
"failed": 0,
"overall_pct": 80
},
"verdict": "AI-READY",
"remediation": [
{
"priority": 1,
"check": "1.4",
"action": "Add h1 tags to remaining 2 landing pages"
},
{
"priority": 2,
"check": "2.3",
"action": "Add /llms-full.txt (full-context version for RAG pipelines)"
},
{
"priority": 3,
"check": "2.6",
"action": "Add Person schema markup to all blog post authors"
}
]
}Step 4 of 5 — The moat: Claude calls the AVR self-audit as an MCP tool. Watch the JSON-RPC handshake, tools/call dispatch, progress notification, and final result.
Step 5 of 5 — The audit result in the dashboard: section scores, remediation priorities, and the AI-READY verdict card.