What I learned shipping SEO for a multilingual static site in 2026
We just launched a travel content site and went through the whole new-domain indexing grind: Google trust window, Bing API quirks, IndexNow, structured data, llms.txt. Here are the notes, because a few things surprised m
We just launched a travel content site and went through the whole new-domain indexing grind: Google trust window, Bing API quirks, IndexNow, structured data, llms.txt. Here are the notes, because a few things surprised me.
The setup
Static multilingual landing + 17 guide pages, English content for travelers: https://fasttrack.vieschool.com/
Nothing exotic: HTML pages, sitemap.xml, schema.org markup, hosted on Vercel.
1. Search Console does not tell you about quality issues
Every page passed technical checks. INDEXING_ALLOWED, fetch SUCCESSFUL, canonical clean. Yet all 17 pages sat at "Crawled - currently not indexed" for weeks. That status is not an error code - it is Google's quality/trust assessment on a new domain, and the only fix is time + crawl pressure + external links. There is no checklist to complete.
2. Bing is actually the fast lane
- IndexNow submissions return 200 and trigger real recrawls
- URL Submission in Webmaster Tools works (route is /webmasters/submiturl?siteUrl=...)
- Request indexing per URL worked and 3 of our pages got indexed within days, e.g. https://fasttrack.vieschool.com/vietnam-visa-on-arrival-vs-evisa.html
- Site Scan is a real SEO audit - it caught meta descriptions over 160 chars and titles over 60 that Google never complained about
3. Watch out: sitemap scope vs website scope in Site Scan
If your sitemap lives in a subfolder like /guides/sitemap.xml, Bing Site Scan in sitemap mode may scan 0 pages silently. Website scope works but got blocked by our host's bot protection (403 + x-vercel-mitigated: deny) since the scan crawler does not use Bingbot IPs.
4. llms.txt is free coverage
One static file at the root and AI crawlers get a clean product summary. Ours: https://fasttrack.vieschool.com/llms.txt - costs nothing, worth doing early.
5. Measure meta length in characters, not bytes
Vietnamese text wrecked our audits until we switched to [...str].length in JS. Bash ${#var} counts UTF-8 bytes and reads ~1.5x too high on accented text.
6. The boring stuff is the bottleneck
Sitemaps, canonical, IndexNow - all done in a day. The actual wait is Google deciding your new domain is trustworthy, which is a 2-4 week crawl pattern. The only thing that moves it is crawl pressure from external surfaces: we blog on Blogger (https://vieschool.blogspot.com/) and Medium with deep links to individual guide pages, not just the homepage.
If you are launching something similar, budget the trust window into your plan. The technical checklist ends fast; the credibility clock does not.
Originally published by Dev.to WebDev. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.