Your next customer probably doesn’t speak your language
Over 75% of online buyers prefer purchasing in their native language, even if they speak English. That’s not a statistic you can ignore if you’re running a store, a SaaS, a blog, or any kind of content-driven site. The question isn’t whether to go multilingual — it’s how to do it without turning it into a second full-time job.
The barrier used to be real: either you paid a translation agency by the word, or you spent hours doing it yourself. Both approaches punish growth. Every new page, product, and post you add just means more translation work waiting for you. That loop ends when you set up an automated pipeline — and that’s exactly what this guide covers.
Manual translation kills momentum
Every time you publish something new, there’s a growing pile of translation jobs waiting. Most site owners give up maintaining multilingual consistency within a few months.
Agencies charge per word — forever
Professional translation scales with your content. More pages, more products, more updates — more invoices. There’s no finish line with per-word billing.
AI removes the ceiling entirely
When translation is automated through your own AI API, it doesn’t matter if you publish 10 pages or 10,000. The cost stays near-flat and the process stays invisible to you.
The three-piece stack that makes this work
You don’t need a complex setup. Three components, properly connected, give you a multilingual WordPress site that runs itself. Here’s what each piece does and why you need all three.
The result is a pipeline where content flows from your editor → to WPML → to the AI → back into your site, without you touching anything in between. You can see exactly how the NEXU addon works here.
One plugin. The whole pipeline connected.
The NEXU AI Translation Addon installs like any WordPress plugin. Once activated, it appears inside your WPML settings and registers as an automatic translator. From that point on, every translation job WPML creates gets handled by the AI engine you’ve configured — automatically, in the background.
Building the automated pipeline — step by step
This is the actual setup process. No fluff. By the end of these steps, your site will be translating content automatically every time you publish.
Install WPML and add your languages
Start with WPML active on your site. Go to WPML → Languages and add every language you want your site to support. This sets up the URL structure and content architecture everything else builds on.
Get your AI API key
Create an account with your preferred AI provider — OpenAI, Anthropic, Mistral, or xAI. Generate an API key. This is what gives the addon access to translate your content. You’ll be billed by that provider at their token rates — not by a translation service with markup on top.
Install the NEXU AI Addon and paste your key
Upload and activate the NEXU AI Translation Addon. In its settings panel, paste your API key and choose your AI model. That’s it for configuration. The addon registers itself as a translator inside WPML automatically.
Assign the AI as your default translator in WPML
In WPML’s Translation Management settings, assign the NEXU AI as the translator for your language pairs. Once this is set, every new translation job WPML generates will automatically be picked up and processed by the AI without any input from you.
Bulk translate your existing content
Use the addon’s bulk translation panel to queue up all your existing posts, pages, products, and templates. Send them to the translation queue and watch the dashboard process everything. From this point forward, new content you publish goes through the same pipeline automatically.
Step 3 in practice — the settings panel
The configuration is about as simple as it gets. Paste your API key, choose your model, set a chunk size for processing long content, and save. That’s the entire setup. Everything after that is automatic.
Step 4 — assign the AI inside WPML
The addon shows up inside WPML’s translator list like any human translator would. You assign it to your language pairs once, and WPML routes all new jobs to it from that point forward. No recurring setup, no checking back in.
Bulk translate — then watch the queue run itself
Select any content type — posts, products, pages, categories, Elementor templates — and send them to the queue. The addon picks them up batch by batch and processes them in the background. You don’t need to stay on the page or keep refreshing.
Add content to jobs and track everything
Need to add specific posts or products to the translation queue without going into each item individually? The addon gives you a dedicated admin page for that. And once translations are running, the statistics panel gives you a clear view of volume, language breakdown, and API usage — so you always know what’s happening.
What the output actually looks like
The thing that surprises most people is how natural the translations read. This isn’t the stilted, awkward output you get from basic machine translation. The AI preserves your writing style, your formatting, and your page structure — in every language. Here’s an English source page and its Hindi translation side by side.
Why this approach is different from the alternatives
There are other ways to handle multilingual WordPress. Here’s an honest comparison of what’s actually different about connecting your own AI API through the NEXU addon.
Expensive, slow, and don’t scale. Fine for marketing copy you need perfect. Not workable for hundreds of product descriptions or frequent content updates.
Convenient but expensive at scale. You lose control over which AI model is used, and you’re paying a margin on top of the actual AI provider’s cost.
Decent for plain text, but breaks complex layouts. No control over model quality, and WPML’s native integration with them is limited and fragile for Elementor or WooCommerce content.
Your API key, your chosen model, direct token-rate billing, deep WPML integration, Elementor and WooCommerce support, real-time queue monitoring, and a cost that doesn’t grow with your content.
A few things worth knowing before you start
Match the AI model to the content type
For high-volume, simple content like product titles and short descriptions, a cheaper model like Mistral works well. For long-form blog posts or pages where tone consistency matters, Claude or GPT-4 gives noticeably better results.
Start with your most important content
When doing the initial bulk translation, prioritize your product pages, homepage, and high-traffic posts first. These are the pages where translation quality has the most direct impact on conversions and SEO.
Use WPML’s review mode for key languages
For languages where you have someone who can review — even occasionally — enable WPML’s pending review status. This lets you publish translations automatically while keeping a human quality check in the loop for your most important markets.
Watch the statistics panel when you first run
The first time you do a bulk run, keep an eye on the stats. You’ll see token consumption and translation volume in real time. This helps you understand your actual costs and adjust model selection or batch sizes if needed.
Look, I'll be honest I was really hoping this would solve my multilingual site headaches, but setup was way more complicated than the guide suggested. The "three piece stack" they talk about? Not as simple as it sounds. I run a recipe blog with WooCommerce, and while the AI translations handle static pages okay, keeping product descriptions and dynamic content updated across languages still feels like a part time job. the old method cost more, but at least I knew what I was getting
Set it and forget it saves money
Got this guide for our company site. the automated translation setup actually cuts down the busywork.