LLMready API and MCP server

One call, no API key and no account: get a site's AI search readiness score, grade, category scores, which AI crawlers its robots.txt allows or blocks, whether it has an llms.txt, and the top 3 fixes. It uses the same live crawl and scoring as a full audit, so the score matches — it just doesn't save a report.

curl "https://answer-engine-ready.lovable.app/api/public/v1/quick-check?url=example.com"

curl -X POST https://answer-engine-ready.lovable.app/api/public/v1/quick-check \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com"}'

const res = await fetch("https://answer-engine-ready.lovable.app/api/public/v1/quick-check?url=" + encodeURIComponent("example.com"));
const check = await res.json();
if (!res.ok) throw new Error(check.message); // 429: wait Retry-After seconds
console.log(check.score, check.grade, check.topFixes[0]?.fix);

{
  "url": "https://example.com/",
  "finalUrl": "https://example.com/",
  "domain": "example.com",
  "score": 74,
  "grade": "C",
  "verified": false,
  "measuredAt": "2026-10-30T09:12:44.120Z",
  "cached": false,
  "categories": [
    { "key": "bot_access", "label": "Bot access", "score": 92, "weight": 0.3 },
    { "key": "extractability", "label": "Content extractability", "score": 68, "weight": 0.25 },
    { "key": "metadata", "label": "Semantic metadata", "score": 61, "weight": 0.2 },
    { "key": "citation", "label": "Citation readiness", "score": 70, "weight": 0.25 }
  ],
  "topFixes": [
    { "label": "Publish /llms.txt", "category": "citation", "impact": 6, "fix": "Add a Markdown index of your key pages at /llms.txt…" }
  ],
  "bots": [ { "name": "GPTBot", "operator": "OpenAI", "purpose": "Model training", "status": "allowed" }, … ],
  "llmsTxt": { "found": false, "url": "https://example.com/llms.txt", "valid": null },
  "llmsFullTxt": { "found": false, "url": "https://example.com/llms-full.txt" },
  "links": {
    "fullAudit": "https://answer-engine-ready.lovable.app/scan?url=https%3A%2F%2Fexample.com%2F",
    "docs": "https://answer-engine-ready.lovable.app/docs#quick-check"
  },
  "limits": { "perMinute": 10, "perDay": 100, "remainingToday": 99 },
  "note": "A score measures how readable a site is for AI crawlers. It does not guarantee that any AI engine will cite the site."
}

• Limits per IP: 10 checks a minute and 100 a rolling day. Responses carry X-RateLimit-Remaining; a 429 carries Retry-After.
• Checking the same URL again within 10 minutes returns the last result with "cached": true, so sites aren't crawled over and over.
• Need every failing check, a shareable report page, the badge or a generated llms.txt? Send people to links.fullAudit, or use the Pro API below.
• In MCP clients the same check is the quick_check tool, also without signing in. Anonymous calls share this per-IP allowance; signed-in accounts get their own.

Base URL https://answer-engine-ready.lovable.app/api/public/v1. Create a key in your dashboard, export it as LLMREADY_API_KEY, and audit a URL:

curl -X POST https://answer-engine-ready.lovable.app/api/public/v1/audits \
  -H "Authorization: Bearer $LLMREADY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com"}'

const res = await fetch("https://answer-engine-ready.lovable.app/api/public/v1/audits", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.LLMREADY_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({ url: "https://example.com" }),
});
if (!res.ok) throw new Error((await res.json()).message);
const audit = await res.json();
console.log(audit.score, audit.grade, audit.fixes[0]?.fix);

import os, requests

res = requests.post(
    "https://answer-engine-ready.lovable.app/api/public/v1/audits",
    headers={"Authorization": f"Bearer {os.environ['LLMREADY_API_KEY']}"},
    json={"url": "https://example.com"},
    timeout=60,
)
res.raise_for_status()
audit = res.json()
print(audit["score"], audit["grade"], audit["links"]["report"])

Send your key as Authorization: Bearer lr_… (or x-api-key). Keys belong to your account, can be revoked anytime and are stored only as a hash — we can't show a key again after you create it.

The MCP server doesn't use keys. quick_check needs nothing; for the other tools your assistant signs you in with OAuth and acts as you.

The API and MCP server are included with Pro (€19/month or €149/year): 1,000 credits per calendar month (UTC), shared between REST and MCP. Failed calls never use credits. Every response carries X-Credits-Remaining. The free quick check never uses credits.

Rate limit: 30 requests per minute per key (per account for MCP).

Errors are JSON: { "error": "code", "message": "…" }. Codes are stable; messages are safe to show to users.

Give ChatGPT, Claude, Cursor or any MCP client the same tools: quick_check, audit_url, get_audit, list_audits, generate_llms_txt, validate_llms_txt and get_usage.

• quick_check works without signing in — 10 checks a minute and 100 a day per IP, shared with the free REST quick check. Signed in, the same limits count per account.
• The other tools need your LLMready Pro account. Called without signing in, the server answers 401 with a WWW-Authenticate challenge, which MCP clients use to open the sign-in screen. They use your Pro credits.

• ChatGPT

  Settings → Apps & connectors → Advanced → turn on developer mode, then Create → paste the server URL. Choose OAuth to use the Pro tools and sign in to LLMready when asked; quick_check also works with no authentication.

• Claude (web and desktop)

  Settings → Connectors → Add custom connector → paste the server URL → Connect. quick_check works right away; approve on the LLMready screen when Claude asks you to sign in for the Pro tools.

• Claude Code

  Run the command below. quick_check works right away; use /mcp inside Claude Code to sign in for the Pro tools.

• Cursor

  Add the JSON below to ~/.cursor/mcp.json (or Settings → MCP → Add). quick_check works right away; click Connect to sign in for the Pro tools.

claude mcp add --transport http llmready https://answer-engine-ready.lovable.app/mcp

codex mcp add llmready --url https://answer-engine-ready.lovable.app/mcp

{
  "mcpServers": {
    "llmready": {
      "url": "https://answer-engine-ready.lovable.app/mcp"
    }
  }
}

{
  "servers": {
    "llmready": {
      "type": "http",
      "url": "https://answer-engine-ready.lovable.app/mcp"
    }
  }
}

curl -X POST https://answer-engine-ready.lovable.app/mcp \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -H "MCP-Protocol-Version: 2025-11-25" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"quick_check","arguments":{"url":"example.com"}}}'

Pro sites can forward AI crawler requests from their own servers so the dashboard shows real GPTBot, ClaudeBot and PerplexityBot visits. Each monitored site gets its own ingest key; send batches of up to 500 requests to https://answer-engine-ready.lovable.app/api/public/v1/crawler-hits. Ready-made snippets for Cloudflare Workers, Vercel/Next.js middleware, Express and nginx logs are in each site's Crawler activity tab.

Hits are matched on the user agent, which crawlers declare themselves — so treat counts as what your server saw, not verified identity.

Agents and MCP clients can find everything on this site without reading the docs. All of these are public, read-only and CORS-enabled.

• No key or sign-in? Use the free quick check first — over REST or as the quick_check MCP tool — and point people to links.fullAudit for the full report.
• Machine-readable docs: /openapi.json and /llms.txt.
• Prefer list_audits or get_audit before running a new audit for the same domain on the same day — reads are free.
• Return the report link to users; it's public and shareable, and the badge embed is in links.badgeEmbedHtml.
• A score measures how readable a site is for AI crawlers. It doesn't guarantee citations in any AI engine — don't promise that.