Technical SEO for Arabic websites: RTL, URLs, fonts and rendering

The check that comes before everything

Before any checklist, open your Arabic page in a browser, view the page source, and search it for a sentence from your service description. If it is there, search engines and AI crawlers find it. If it is not, your text is drawn after load by JavaScript, and that is your first technical problem regardless of what follows.

Many recent Gulf sites are built on frameworks that work this way, and look excellent to a visitor and empty to a crawler. The fix is not to abandon the framework but to enable server rendering or pre-generation for the marketing pages.

The shared rules, briefly

Technical SEO for an Arabic site starts from the same rules as any site:

• Crawling and indexing: a robots.txt that does not block what should be indexed, an up-to-date sitemap, and no noindex tag left over from development.
• One URL per page: a correct canonical, and a permanent redirect between www and non-www and between http and https.
• Structure: important pages within three clicks of the homepage, and no orphan pages.
• Speed: real-user data on mobile, detailed in the speed article.

What follows is what Arabic adds on top.

Item 1: direction through logical properties, not a flipped design

The old way to Arabise an English site: copy the stylesheet and flip every left to right. The result is mirrored icons, numbers that break, margins that drift with every later edit, and a second stylesheet to maintain.

The correct way: dir="rtl" on the Arabic page, and logical CSS properties (margin-inline-start instead of margin-left, text-align: start instead of left) in one stylesheet that works in both directions. Every modern browser supports it, and it is what lets the two versions change together instead of one silently breaking.

Google does not measure RTL directly, but a wrongly mirrored page raises bounces and lowers engagement, and those it does measure.

Item 2: lang and dir where they belong

<html lang="ar" dir="rtl"> on the Arabic page and lang="en" on the English one. It sounds obvious, and I find it wrong on roughly a third of the sites I audit: an Arabic page with lang=“en”, which renders in the wrong font on some systems, confuses the engine about the content language, and trips screen readers.

Numbers and English words inside Arabic text sometimes need explicit isolation with bidirectional control characters to stay readable, especially ranges such as “6–12 months”. Small details, but a Gulf reader sees them immediately.

Item 3: URLs in Arabic or Latin letters

Google indexes Arabic URLs and displays them in Arabic in results, so they are not an error. But they carry a practical cost: copied, they become long percent-encoded strings; they break in some email systems and platforms; they are hard to share.

The choice I recommend for most Gulf sites: short, understandable Latin URLs, in English or a clear transliteration, with a fully Arabic title and content. The URL is not a ranking signal worth sacrificing shareability for. If you choose Arabic, keep it short and stable, and never change it later.

Item 4: Arabic fonts and their weight

This is the item that slows Arabic sites most and appears on no English checklist. A single Arabic font, with its connected letterforms, can weigh twice its Latin equivalent, and the typical site loads four weights of it on every page before any text appears.

The rules:

• Two weights in most cases: regular and bold.
• WOFF2 only, with unused scripts subset out of the file.
• font-display: swap so text shows in a system font until the web font arrives.
• Preload only the font that draws the main heading.
• A budget: no more than about 60 KB per weight.

Done properly, LCP on mobile improves immediately, because the main heading no longer waits for a heavy font file.

Item 5: hreflang and the two versions

Every Arabic page points to its English counterpart and to itself, and vice versa, with an x-default. Without it Google picks one version and hides the other. The full detail is in the hreflang guide; here it is enough to say it is the most repeated technical error on Gulf sites.

Item 6: structured data in Arabic

Schema markup works in Arabic as it does in English, and its text values are written in the language of the page. The common mistake is copying the English markup unchanged onto the Arabic page, so an Arabic page carries the business name and description in English. One block per page in its language, with shared identifiers kept shared.

The technical checklist for an Arabic site

This site as a sample

The site you are reading is built on these rules: Arabic first, RTL through logical properties, two fonts of limited weight, and all text in the page source. Check it with any tool you trust before paying anything. In the Digital Presence Audit the technical diagnosis of the Arabic version is a fixed item, not an appendix to the English check.