QQQ AI Visibility Checker
Check how visible a company website is to AI search with the free QQQ AI Visibility Checker (qqq.ee). Runs a technical AI readiness check of a domain (AI crawler access in robots.t…
Meelis
@meelis545
Install
$ openclaw skills install @meelis545/qqq-ai-visibilityQQQ AI Visibility Checker
Free tool by QQQ, an SEO and AI search agency in Estonia. It shows three things separately, without a combined score:
- Mentions: do AI assistants with web search (ChatGPT, Gemini, Perplexity, Claude) name the company when buyers ask about its field?
- Citations: do those answers cite the company's own website as a source?
- Technical readiness: can AI crawlers reach and read the website?
Web tool: QQQ AI Visibility Checker (Estonian: AI nähtavuse kontroll). Full reports are public pages: all reports.
What this skill touches
- Network:
https://qqq.ee/mcponly. Use exactlyqqq.ee, notwww.qqq.ee(www answers with a 301 redirect, which breaks the POST). - Files: none. Tools:
curl. Nothing is installed. - No API key, no account, no environment variable.
Setup
Nothing to set up. The API is public and free, with a daily limit (see Limits).
How to call it
One endpoint, JSON-RPC 2.0 over HTTP POST. Every call has the same shape; only the tool name and the arguments change:
curl -s https://qqq.ee/mcp \
-H 'content-type: application/json' \
-H 'accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"TOOL","arguments":{ARGS}}}'
The answer is JSON:
result.content[0].text: a ready-to-show text summary with links. Show this to the user.result.structuredContent: the same data as fields, for your own reasoning.result.isError: true: the domain was refused or the limit is reached. The text says why and still gives a link the user can open.
language is "en" (default) or "et" (Estonian). Use "et" when the user writes in Estonian.
Endpoints (tools)
check_ai_visibility: technical check of a domain
curl -s https://qqq.ee/mcp \
-H 'content-type: application/json' \
-H 'accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"check_ai_visibility","arguments":{"domain":"example.com","language":"en"}}}'
domain(required): a public domain or URL, for exampleexample.comorhttps://www.example.com/page.- Fetches the website's own pages (homepage, robots.txt, sitemap, llms.txt): HTTPS, AI crawlers allowed (GPTBot, ClaudeBot, PerplexityBot and others), Googlebot allowed, XML sitemap, title, meta description, one H1, canonical, noindex, lang, Organization/Person structured data, content in HTML (not JS-only), og:image, key pages.
- Returns the passed and failed checks, the top 3 fixes, the latest public AI answer report numbers if one exists, and a prefilled link to the full AI answer check on qqq.ee.
structuredContentfields:readiness(passed,total,checks[]withidandstatusok/missing/unknown),blockedAiCrawlers,topFixes,latestReport,links.fullCheck,links.report.- It does not ask any AI model. Asking ChatGPT, Gemini and Perplexity about the company happens only on qqq.ee, behind bot protection. Send the user the
links.fullChecklink for that.
get_public_report: the latest public AI answer report of a domain
curl -s https://qqq.ee/mcp \
-H 'content-type: application/json' \
-H 'accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"get_public_report","arguments":{"domain":"example.com","language":"en"}}}'
- Reads the stored report only: no website is fetched and no AI model is asked.
- Returns the report date, market and field, in how many search-grounded AI answers the company was named, how often its own website was cited, a summary per AI engine, the questions asked, the cited source domains, other companies named, and short answer excerpts.
- Without a report it says so and gives the prefilled link to run the full check on qqq.ee.
get_tool_info: what the tool is, limits and links
curl -s https://qqq.ee/mcp \
-H 'content-type: application/json' \
-H 'accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"get_tool_info","arguments":{"language":"en"}}}'
Workflow
- User asks about a domain's AI visibility: call
check_ai_visibility. - If the text has a "Latest full report" line, or the user asks what ChatGPT, Gemini or Perplexity said, call
get_public_reportfor the details. - Show the result text. Keep the links in it exactly as they are and make them clickable, so the user can open the report or start the full check.
- If there is no report, tell the user that the AI answer check runs on qqq.ee and give the
links.fullChecklink (QQQ AI Visibility Checker if you have no domain link).
Rules
- Mentions, citations and technical readiness are separate results. Do not add them up into one score.
- Do not invent AI answers or numbers. Only report what the API returned, with its date.
- A technical check passing does not mean AI assistants mention the company. Say which of the three you actually measured.
- Do not retry in a loop. On
isErroror a limit message, show the message and the link. - Do not call it for private or internal hosts, IP addresses or
localhost. The API refuses them anyway. - Keep the attribution line "Free AI visibility checker by QQQ (qqq.ee)" when you quote a result.
Limits
| Tool | Limit |
|---|---|
check_ai_visibility | 5 per day per IP address, 300 per day in total. An invalid domain does not use an attempt. |
get_public_report | no limit |
get_tool_info | no limit |
The same server works as a remote MCP server at https://qqq.ee/mcp (Streamable HTTP, no auth). Setup for Claude, ChatGPT, Cursor and VS Code: QQQ MCP server docs.
Top skills in this category
ATXP
@emilioaccAccess ATXP paid API tools for web search, AI image generation, music creation, video generation, X/Twitter search, email, and agent account management. Use...
Session-logs
@guogang1024Search and analyze your own session logs (older/parent conversations) using jq.
Reddit (read only - no auth)
@buksan1950Browse and search Reddit in read-only mode using public JSON endpoints. Use when the user asks to browse subreddits, search for posts by topic, inspect comment threads, or build a shortlist of links to review and reply to manually.
Computer Use
@ram-raghav-sFull desktop computer use for headless Linux servers. Xvfb + XFCE virtual desktop with xdotool automation. 17 actions (click, type, scroll, screenshot, drag,...
tech-news-digest
@dinsteinGenerate tech news digests with unified source model, quality scoring, and multi-format output. Six-source data collection from RSS feeds, Twitter/X KOLs, Gi...