# Whitelist policy: only the homepage, contact page, policy pages and
# spotlight pages (plus the Rendering Photos they embed) may be crawled.
# Everything else (designer app, app.js, pricing data, API routes, other
# assets) is off-limits to crawlers.
User-agent: *
Allow: /$
# Kept crawlable on purpose: /index.html now 301s to / (vercel.json). Crawlers
# must be able to fetch it to see the redirect and drop the duplicate URL.
Allow: /index.html
Allow: /contact.html
Allow: /kitchen-cabinets-aurora.html
Allow: /spotlight.html
Allow: /spotlight/
Allow: /learn.html
# Free public .kit viewer (kit.html + the two scripts it runs on; nothing else under /kit*).
Allow: /kit$
Allow: /kit/zh$
Allow: /kit/fr$
Allow: /kit.html
Allow: /kit-zh.html
Allow: /kit-fr.html
Allow: /kit-viewer.js
Allow: /kit2020-import.js
# Free public PDF editor (the three language pages, its script and stylesheet,
# and the two vendored engines it lazy-loads). Google renders a page before
# judging it, so the CSS/JS have to be fetchable or /pdf grades as an unstyled
# wall of text — the same trap that bit tw.css and spotlights.json.
Allow: /pdf$
Allow: /pdf/zh$
Allow: /pdf/fr$
Allow: /pdf.html
Allow: /pdf-zh.html
Allow: /pdf-fr.html
Allow: /pdf-editor.js
Allow: /pdf-editor.css
Allow: /assets/pdfjs/
Allow: /assets/pdf-lib/
Allow: /learn/guide/
Allow: /community.html
Allow: /privacy.html
Allow: /terms.html
Allow: /Rendering%20Photos/

# MUST stay crawlable: the public pages' entire stylesheet. Google and Bing
# render a page before judging it — block the CSS and they see an unstyled
# wall of text (and flag it mobile-unfriendly). This replaced the old
# cdn.tailwindcss.com script, which was external and so never robots-gated.
Allow: /assets/tw.css

# llms.txt — the plain-Markdown site guide written for AI assistants
# (ChatGPT, Gemini, Claude, Perplexity). It returned 200 but the catch-all
# Disallow below made every compliant AI crawler skip it, so the one file we
# wrote specifically for them was the one file they could not read
# (found 2026-08-25 — fifth recurrence of the whitelist trap).
Allow: /llms.txt

# IndexNow / site-verification key file (Bing Webmaster Tools).
Allow: /71abb468be9141e38aebbf3d1f7283ae.txt

# RSS feed（2026-08-29 新增）。⭐ 白名单制的老规矩：新增任何公开文件都必须在
# 这里点名放行，否则文件末尾那条 catch-all `Disallow: /` 会把它挡死 —— llms.txt、
# sitemap.xml、spotlights.json、tw.css、IndexNow key 已经各栽过一次。
Allow: /feed.xml

# The sitemap itself. Google/Bing fetch it under robots rules too — without
# this line the catch-all Disallow below makes GSC report "Couldn't fetch".
Allow: /sitemap.xml

# spotlight.html and the homepage trends section render their article grid by
# fetching this JSON. Google's renderer honors robots for subresources too —
# with it blocked, spotlight.html rendered as an empty shell and GSC live test
# rejected it as a Soft 404 (found 2026-08-23).
Allow: /spotlights.json

# community.html builds its gallery/Q&A/challenges/pros grids and learn.html
# builds its video grid from these endpoints (all public, no auth). Blocked,
# the crawler's renderer sees empty grids + "No public designs yet" — the same
# Soft-404 class as spotlights.json above (found 2026-08-28, recurrence #6).
Allow: /api/community-
Allow: /community/inspiration.json
Allow: /community/answers-index.json
# The redesigned community.html additionally renders door and countertop swatch
# strips/tiles whose <img> URLs are built at runtime from the colour catalogs.
# Same trap as the JSON above: blocked, Google's renderer sees broken images and
# grades the page a Soft 404 (recurrence #7). The real folder is "countertop
# colors" with a literal space, percent-encoded here like /Rendering%20Photos/.
Allow: /door-colors/
Allow: /countertop%20colors/

# ⭐ 白名单第 8 次栽跟头 (2026-09-05 全站体检 ACQ-001)。上面只放行了图片文件夹,
# 却没放行「生成这些图片 URL 的脚本」——community.html 里 /i18n.js 被挡 =>
# 页面脚本抛 "T is not a function" 后整块中止, Googlebot 拿到的正文从 6,573 字
# 掉到 822 字、<img> 从 16 张掉到 1 张 (learn.html 少掉视频宫格, index.html 少掉
# 整套 i18n 文案)。以下五个都是公开页首屏渲染必需的**数据/文案**脚本, 逐个查过
# 没有任何报价数据 (grep '$'/CAD/USD: door-colors*.js、countertop-colors.js、
# page-i18n.js 全 0 命中; i18n.js 只有 PayPal 最低充值和预算区间这两句 UI 文案,
# 本来就印在页面上)。⛔ COMPLETE_PRICING.js 依旧不放行, 报价数据库永远不进白名单。
Allow: /i18n.js
Allow: /page-i18n.js
Allow: /door-colors.js
Allow: /door-colors-extra.js
Allow: /countertop-colors.js

# Emma 的头像图（index.html 的 <picture> 用 webp + png 两份，learn.html 用 png）。
# 同样是渲染时可见的图片：挡住 Google 就在首页渲染里看到一张裂图 (2026-09-05)。
Allow: /avatar.webp
Allow: /avatar.png
Allow: /avatar-full.webp
Allow: /avatar-full.png

# 干净 URL (2026-09-05 ACQ-010, vercel.json 新增 rewrite): /learn、/community、
# /spotlight 现在返回和对应 .html 完全相同的页面, 而那三个页面都自带指向 .html
# 的 canonical, 所以放行它们不会产生重复收录, 只是让分享出去的短链接不至于撞上
# 下面那条 catch-all。⛔ /designer 和 /start 这两条新 rewrite **故意不放行**:
# 设计器引擎不进白名单 (Kim 的老规矩), catch-all 已经把它们挡住。
Allow: /learn$
Allow: /community$
Allow: /spotlight$

# 新增公开页 (2026-09-05): /for-pros、/about、/warranty —— 两种写法都点名,
# 干净 URL 走 vercel.json 的 rewrite, .html 是自 canonical 的那一份。
Allow: /for-pros$
Allow: /for-pros.html
Allow: /about$
Allow: /about.html
Allow: /warranty$
Allow: /warranty.html

# Brand icons: Google only shows a site's favicon in search results if it can
# crawl it. Blocked, every SERP entry shows the generic globe (found 2026-08-28).
Allow: /brand/

# In-article images for the guide pages (install steps, designer screenshots).
# Blocked, they are invisible to Google Images and missing from the rendered
# page Google evaluates (found 2026-08-28). AI-training use is already handled
# by the X-Robots-Tag: noimageai response header, which needs the fetch allowed.
Allow: /learn/assets/

# The designer tool and its 2D/3D editor interfaces (and the JS that powers
# them) must never be crawled or scraped. These are named explicitly below
# for clarity; the catch-all Disallow at the end already blocks everything
# outside the whitelist above.
Disallow: /designer.html
Disallow: /start.html
Disallow: /settings.html
Disallow: /app.js
Disallow: /edit-mode-3d.js
Disallow: /design-helper-3d.js
Disallow: /countertop-edit.js

# ⭐⭐ 内部页面**故意一个都不在这里点名**。robots.txt 是公开可读的：在这里
# 写 `Disallow: /<路径>` 等于对所有人发布一份「这个站有哪些内部页面」的目录
# (Kim 2026-08-14「不希望这个链接可以被找到」)。下面那条 catch-all 已经把
# 白名单以外的一切都挡住了，效果完全一样，只是不再点名。内部页面另有
# <meta name="robots" content="noindex,nofollow"> 和 vercel.json 里的
# X-Robots-Tag 两道，都不依赖这里点名。
#
# ⚠️ 所以：**以后新增内部页面时，不要往这个文件里加 Disallow 行。** 加了
# 就是把它公布出去。
#
# Catch-all: block everything not explicitly allowed above.
Disallow: /

Sitemap: https://www.flipkitchen.ca/sitemap.xml