win777.game looks like a gambling or betting site, is ranked in the top 10,000 sites and was first seen by the web archive in 2024. It uses Cloudflare, Cloudflare DNS and Google Tag Manager; we found no traces of an AI site builder.
The page shows little text before JavaScript runs. Sites drawn by JavaScript show fewer signals until we can render them in a browser.
We look for what AI site builders such as Lovable, Bolt and v0 leave in a page. Code written with an AI assistant such as Cursor, Claude Code or ChatGPT leaves nothing to find, so we can’t tell whether AI helped.
AI traces: None found
How we know
AI site builders leave traces in the sites they make. We look for the traces of 53 of them and show every match with the line it came from, so you can check it yourself.
We test this on 2,275 sites whose makers we know, and publish what we get wrong.
Prompt for your coding assistant (Claude Code, Cursor, your builder’s chat)
The home page of https://win777.game sends only 59 characters of text before its scripts run, so AI crawlers that don’t run JavaScript see an almost empty page. Make the server send the page’s content in its HTML: render it on the server, or prerender each page at build time (static generation), so its headings, text and links are in the HTML before any script runs. Keep the interactive parts working as they do now.
+25points
Sitemap
Now None found.
Publish an XML sitemap and list it in robots.txt.
Details
Why it matters: A sitemap lists every page, so crawlers find the ones no link leads to.
Prompt for your coding assistant (Claude Code, Cursor, your builder’s chat)
Add an XML sitemap at https://win777.game/sitemap.xml that lists the site’s public pages, in the sitemaps.org format, and add the line "Sitemap: https://win777.game/sitemap.xml" to https://win777.game/robots.txt. Generate it from the site’s routes, so new pages are added to it.
+10points
Optional: 4 fixes few sites have yetup to +18 points
Files for AI agents
Now None of the 8 we look for.
If agents could use the site, publish what they need where they look.
Details
Why it matters: Agents act for people: they look for these files to know what a site lets them do. Worth it only if agents can use the site.
Prompt for your coding assistant (Claude Code, Cursor, your builder’s chat)
If AI agents could use https://win777.game for its visitors (to search it, book, buy or sign in), publish what they need at the standard addresses, and only what the site really offers: an MCP server card at /.well-known/mcp/server-card.json if it runs an MCP server, an API catalog at /.well-known/api-catalog (RFC 9727) listing its public APIs, and OAuth protected resource metadata at /.well-known/oauth-protected-resource (RFC 9728) if agents sign in.
+10optional
llms.txt
Now None found.
Publish a short Markdown summary of the site at /llms.txt.
Details
Why it matters: A short summary for AI models, at an address they can find. A proposal: few AI companies say they read it yet.
Prompt for your coding assistant (Claude Code, Cursor, your builder’s chat)
Create https://win777.game/llms.txt as llmstxt.org describes it: a Markdown file with an H1 naming the site, a one-paragraph summary in a blockquote, and short sections listing its most useful pages as links, each with a line on what it holds. Serve it as text/plain or text/markdown.
+4optional
Links for AI tools
Now None in the home page’s headers.
Point AI tools to the site’s files in the home page’s Link header.
Details
Why it matters: A Link header points AI tools to a site’s files from any page, without guessing addresses.
Prompt for your coding assistant (Claude Code, Cursor, your builder’s chat)
Make the home page of https://win777.game send a Link header (RFC 8288) that points AI tools to the site’s files, for example: Link: </sitemap.xml>; rel="sitemap", </llms.txt>; rel="describedby"; type="text/plain". List only files the site serves.
+2optional
Markdown copy
Now Not offered: it sends the page’s HTML.
Answer requests for Markdown with a Markdown version of the page.
Details
Why it matters: A Markdown copy gives AI tools the page’s words without the HTML around them.
Prompt for your coding assistant (Claude Code, Cursor, your builder’s chat)
When a request to https://win777.game asks for Markdown (its Accept header includes text/markdown), answer with a Markdown version of the page’s main content and the header Content-Type: text/markdown. Requests for HTML keep getting HTML; add Vary: Accept so caches keep the two apart.
+2optional
Every check, with its evidence
Can AI find it?
robots.txtFound, 680 linesFound
SitemapNone foundNot found
Links for AI toolsNone in the home page’s headersNot found
Search listingAllowed: no noindex ruleAllowed
Can AI read it?
Text without scriptsLittle, 59 characters: scripts draw the restNot found
Keeping AI out is the site’s choice, so it never lowers the score.
AI trainingOpenAI, Anthropic, Google, Apple and 2 more kept out, Meta and ByteDance partlyPartly
AI answers and searchGoogle kept out; the rest let inPartly
Every AI crawler, and how we checked
AI companies
CompanyTo train its modelsTo answer people and search
OpenAIChatGPTTo train its modelsKept outGPTBotline 56To answer people and searchLet inOAI-SearchBotChatGPT-User
AnthropicClaudeTo train its modelsKept outClaudeBotline 47anthropic-ailine 62Claude-Webline 65To answer people and searchLet inClaude-SearchBotClaude-User
GoogleGeminiTo train its modelsKept outGoogle-Extendedline 53To answer people and searchKept outGoogle-CloudVertexBotline 90
PerplexityTo train its modelsNone namedTo answer people and searchLet inPerplexityBotPerplexity-User
MetaMeta AITo train its modelsPartlymeta-externalagentline 74, kept outFacebookBotlet inTo answer people and searchLet inmeta-externalfetchermeta-webindexer
AppleApple IntelligenceTo train its modelsKept outApplebot-Extendedline 38To answer people and searchNone named
AmazonTo train its modelsKept outAmazonbotline 35To answer people and searchLet inAmzn-SearchBotAmzn-User
Common Crawlopen data for trainingTo train its modelsKept outCCBotline 44To answer people and searchNone named
ByteDanceTo train its modelsPartlyBytespiderline 41, kept outTikTokSpiderlet inTo answer people and searchNone named
MistralLe ChatTo train its modelsNone namedTo answer people and searchLet inMistralAI-User
5 of the 27 other AI crawlers we know are kept out: cohere-ai, Diffbot, Omgili, ICC-Crawler and Cotoyogi.
Like 25% of the 1,283 popular sites whose robots.txt we’ve read since 29 Oct 2026, it keeps at least one AI training crawler out. Of the 1,585 popular sites we’ve checked since 29 Oct 2026, 11% have a page drawn by scripts, 3% send a Markdown copy, 10% have an llms.txt and 4% offer a file for AI agents. How many websites block AI crawlers?
We read the site once, as our named bot: its robots.txt, read as crawlers read it, its home page and the files AI tools ask sites for. What we couldn’t read says so. robots.txt asks and doesn’t enforce: a firewall can treat AI crawlers differently from us.
win777.game describes itself: “Play 5000+ casino games at India's most trusted casino. Instant INR withdrawals, 24/7 support, exclusive bonuses. Join 100,000+ daily players!” It looks like a gambling or betting site. See the evidence
How old is win777.game?
The web archive first saw it on 3 Jan 2024, 2 years ago. See the evidence
What is win777.game built with?
It uses Cloudflare, Cloudflare DNS and Google Tag Manager; we found 3 technologies in all. See the evidence
How popular is win777.game?
It is among the top 10,000 sites, by Tranco and Chrome UX Report. See the evidence
Was win777.game made with AI?
We found no traces of an AI site builder. That doesn’t prove a person built it: AI coding assistants leave no traces. See the evidence
Automated findings from the site’s public home page and public records, as of 7 Nov 2026. They describe what we found, not the people or business behind the site, and they may be incomplete or out of date. This is information, not advice. Terms · Report an error
What this scan can’t tell you
Whether the business is real, honest or delivers what it sells.
Whether a site was made with AI when no traces were left: coding assistants and hand edits leave none.
How many people visit: we show popularity bands from public lists, not visits.
Anything beyond the home page as it was at the time of the scan.