# movin.design robots policy. # Search engines → full access (SEO). # AI search / assistants → allowed, so pages can be cited in AI answers # (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, # Claude-User, PerplexityBot fall under "*"). # AI training crawlers → our OWN work only. The clip archive (other # creators' re-hosted clips) stays closed to # training, to protect the creators. # # Split 2026-10-19 (docs/adsense-remediation.md, Phase 4): blocking training # crawlers site-wide also kept every future model from learning about the # components, shaders, tools and writing, which are ours to share. # Google-Extended controls Gemini training AND Gemini app grounding, so it is # opened on the same sections. It never affects Google Search. # # Cloudflare enforces this policy at the edge too (AI bot policies, AI Crawl # Control). Change both together: docs/cloudflare.md. # ── Major AI training crawlers: our own sections only ── # Longest match wins (RFC 9309), so each Allow beats the final "Disallow: /". # "/$" is the home page alone. English paths: enabling a locale means adding # its // copies of these lines (src/lib/locales.mjs, docs/i18n.md). User-agent: GPTBot User-agent: ClaudeBot User-agent: Google-Extended User-agent: CCBot User-agent: Applebot-Extended Allow: /$ Allow: /components/ Allow: /r/ Allow: /shaders/ Allow: /easings/ Allow: /slideshow/ Allow: /icons/ Allow: /blog/ Allow: /about/ Allow: /services/ Allow: /llms.txt Allow: /llms-full.txt Disallow: / # ── Other training / scraping crawlers: blocked ── User-agent: Bytespider User-agent: Amazonbot User-agent: meta-externalagent Disallow: / # ── Everyone else (search engines + AI-search assistants): allowed ── User-agent: * Allow: / Disallow: /410/ Disallow: /saved/ Disallow: /auth/ Disallow: /account/ Disallow: /api/ # Unsubscribe links from our emails (noindex too). Disallow: /email/ # The MCP server and its sign-in pages (docs/mcp-plan.md): for clients, never for crawlers. Disallow: /mcp Disallow: /oauth/ # Internal search results are thin/duplicate by Google's own guidance, and the # query only exists client-side. The pages it lists are all crawlable elsewhere. Disallow: /search # Same three exclusions inside every language prefix (/ja/410/, /ko/search…). # Written as a wildcard rather than one block per locale so adding a language # stays a single edit in src/lib/locales.mjs (see that file's contract). Disallow: /*/410/ Disallow: /*/search Sitemap: https://movin.design/sitemap.xml