Paying per language
Webflow sells localisation as an add-on you attach to a site plan, and it meters two things at once: how many locales you publish and how many words you translate per month. Localize Essential is $9 a month and stops at 3 locales and 10,000 words. Localize Advanced is $29 a month and stops at 10 locales and 50,000 words. Past that, Webflow's pricing page says Custom and points you at a sales call. Those are the live figures on webflow.com/pricing as of today. If you sell into two markets they are cheap enough to ignore. If you are heading toward twelve, the shape of the bill is worth looking at before you get there.
What the meter actually counts
The word cap is the part people miss. Locales are a number you choose once a year. Words are a number your marketing team moves every week, and the 10,000 monthly words on Essential covers a landing page and a handful of CMS entries, not a documentation set. A site that publishes steadily will hit the word ceiling long before it hits the locale ceiling, which pushes you to Advanced for reasons that have nothing to do with how many languages you serve.
There is a second split worth knowing. Localized URLs and automatic visitor routing sit on Advanced, not Essential. So the $9 tier gives you translated content without the two things that make translated content findable: a clean per-market URL and a visitor who lands on their own language without clicking. You can build a locale switcher on Essential, but the visitor has to use it.
None of this is hidden or unfair. Webflow is running machine translation and serving per-locale variants from its own infrastructure, and that costs real money. The point is that the price is indexed to your publishing volume rather than to your engineering effort, and those two numbers drift apart as a company grows.
The same thing in a framework is a config block
Astro's i18n support is a field in astro.config.mjs. You list your locales, name a default, and decide whether the default gets a URL prefix. Folders under src/pages/ matching those locale names become routes. Helpers like getRelativeLocaleUrl() generate correct links, and Astro.preferredLocale reads the browser's Accept-Language header so you can route a visitor without a click. Adding a fourth language is a folder. Adding a twelfth is also a folder.
The work does not vanish, it moves. You still need translations, and translation is the expensive part whether a platform bills you for it or you buy it from a vendor. What changes is that translation becomes a line item you negotiate separately, at whatever volume you actually need, instead of a tier you buy in jumps of 40,000 words. You also stop paying a recurring fee for routing logic that does not change after you write it.
We saw the same pattern on the Stay World Class migration. Moving from Webflow and Xano to Next.js, NestJS and Supabase took Lighthouse from 55.91 to 91 out of 100 and cut LCP by 77%, with a four stage ETL that mapped the legacy IDs into PostgreSQL UUIDs so nothing lost its history. The performance numbers get the attention, but the quieter result was that the recurring per-feature charges turned into code that sits in the repo and bills nothing.
Do the arithmetic for your own market count
Take the number of languages you will publish in twenty four months, not the number you publish today, and multiply the word volume by your actual editorial cadence. Two locales and a slow blog: Localize Essential costs $108 a year and you should stop reading. Six locales and a docs site your team updates weekly: you are on Advanced at $348 a year plus a site plan, and you are watching a word counter that has nothing to do with your business.
Somewhere past ten locales the conversation stops being about price. It becomes about whether the routing rules, the per-market SEO and the translation pipeline are things you can version, test and change on a Tuesday, or things you file a ticket about. That is the real threshold, and it arrives at a different point for every team.
When staying on Webflow is the right answer
If you serve two or three markets and your content changes monthly, Localize Essential at $9 is cheaper than the first hour of engineering time you would spend replacing it, and it will stay cheaper for years. Buy it and go do something else.
The same holds if your team has no developer. Localisation in a framework is simple code, but it is still code, and code with nobody to maintain it is worse than a subscription. A marketing team that can translate a page in the Webflow canvas without filing a pull request is getting real value for $29 a month.
Migrate when the meter starts steering your decisions. When somebody asks whether to publish in Portuguese and the first question back is about the word allowance rather than the market, the pricing model has begun making product choices for you. That is the signal, not the invoice total.
