If you already pay for WPML and you are looking at a 100-page or 500-page site that needs translating into 2 or 3 languages, there are two paths in front of you. Pay for WPML's translation credits, or bring your own AI API key and pay the provider directly. This guide explains the difference, shows the real numbers and walks through a practical step-by-step setup using AutoMLP with OpenAI, Gemini or Chrome built-in AI.
How WPML's built-in AI translation works
WPML has its own Advanced Translation Editor (ATE) with AI translation baked in. You pick a provider from inside WPML, the editor sends your content to that provider through WPML's infrastructure. The cost is billed back to you in "translation credits." One credit roughly equals one translated word. Credits come bundled with your annual WPML plan.
The Multilingual CMS plan includes a yearly allowance that sounds generous on paper. In practice it gets used up faster than people expect. A single 1,500-word blog post translated into 3 languages eats 4,500 credits. A WooCommerce store with 500 products easily burns through a year's allowance before launch. When the bundled credits run out you buy more. The per-credit cost is where this stops being affordable.
Why WPML's AI credits get expensive fast
Here is the cost picture for a typical multilingual project, translating from English into Spanish, French and German:
| Site size | Total words to translate | WPML AI credits (approx) | Your own OpenAI key | Your own Gemini key |
|---|---|---|---|---|
| Small marketing site | 30,000 words × 3 languages | $90 to $110 | $9 to $12 | Under $2 |
| Content blog (200 posts) | 200,000 words × 3 languages | $600+ | $60 to $70 | $8 to $12 |
| WooCommerce store (500 products) | 350,000 words × 3 languages | $1,000+ | $100 to $120 | $14 to $20 |
WPML's credits work out to roughly $1 per 1,000 words once you exhaust the bundled allowance. A direct OpenAI key costs about $0.10 per 1,000 words for the same translation work. Gemini Flash is cheaper still, often under $0.02 per 1,000 words. Chrome built-in AI is free because it runs locally in the browser.
On a small site the credit system is fine. The moment you have WooCommerce, multiple languages or any kind of content volume, the math stops working in WPML's favour. That is the gap AutoMLP fills.
What AutoMLP does differently
AutoMLP is an addon that sits next to WPML and gives you bring-your-own-key AI translation. Instead of paying WPML for credits, you connect your own API key from OpenAI or Google AI Studio, and AutoMLP sends your translation jobs directly to that provider. You pay the provider's raw API rate, no markup, no credit conversion.
AutoMLP supports three AI options today:
- OpenAI for marketing copy, editorial content and anything where tone matters. GPT-4o is the recommended model for production work.
- Google Gemini for high-volume sites that need cost-efficient bulk translation. Gemini 1.5 Flash is roughly 10× cheaper than OpenAI for comparable quality on most language pairs.
- Chrome built-in AI for free in-browser translation. No API key, no provider account. Runs entirely in modern Chrome on your machine.
All three sit inside WPML's existing workflow. Pages, posts, custom post types, WooCommerce products and String Translation entries are all supported. WPML keeps managing the multilingual structure (hreflang, language URLs, post relationships). AutoMLP just replaces the translation engine.
Bring your OpenAI or Gemini API key, or use free Chrome AI. Translate WPML sites in bulk at provider cost, often 10× cheaper than WPML credits.
Step by step: how to translate a WPML site with AutoMLP
The full setup takes about ten minutes if it is your first time. Once it is connected, every future translation job is a few clicks from the standard WPML pages and posts screens.
Step 1. Install AutoMLP alongside WPML
Make sure WPML Multilingual CMS is already installed and your languages are configured. Then install AutoMLP from your purchase area and activate it on the same WordPress site. AutoMLP will detect WPML automatically and add its own settings panel under the WordPress admin menu.
Step 2. Get your AI API key (or pick Chrome AI for free)
Pick the provider that fits your workload:
- For OpenAI: sign up at platform.openai.com/api-keys, add a payment method and create a secret key. Copy it to a safe spot before you leave the page.
- For Gemini: visit aistudio.google.com/apikey, sign in with a Google account and create a key. The Gemini free tier is generous and usually covers small projects with no billing setup.
- For Chrome built-in AI: no key needed. Just make sure you are using a recent version of Chrome with the Translation API enabled.
Step 3. Connect the key in AutoMLP settings
Open the AutoMLP settings page in your WordPress admin. Select your AI provider from the dropdown, paste your API key into the field and save. AutoMLP runs a quick connection test so you know the key is working before you start translating.
Step 4. Open the bulk translation panel
Go to Pages or Posts in your WordPress admin. AutoMLP adds a new column and bulk action specifically for AI translation. You will see translation status indicators next to each item showing which languages already have a translation and which still need one.
Step 5. Select the items you want to translate
Tick the checkbox next to each post or page, or use the "Select all" checkbox at the top to grab everything on the current screen. For very large sites you can also filter by category, author or date first to translate in smaller batches.
Step 6. Click the bulk translate button
Pick your target language (or all configured languages at once) from the bulk action dropdown, then click Bulk Translate. AutoMLP queues the jobs and processes them in the background, talking to your selected AI provider directly. You can keep working on other things while the queue runs.
Want a deeper walkthrough on the bulk side specifically? The WPML bulk translation guide covers batching, queue management and how to handle rate limits cleanly.
Which AutoMLP provider should you pick?
All three providers work with the same workflow. The choice comes down to what kind of content you are translating and how much volume you have:
- Pick OpenAI if quality and tone matter more than cost. Best for marketing sites, brand-sensitive content and customer-facing copy where the translation will be read carefully. GPT-4o handles idiom and context better than the cheaper options.
- Pick Gemini if you have a lot of volume and budget matters. Gemini Flash keeps quality close enough to OpenAI for most language pairs, at a fraction of the cost. Good for content blogs, large WooCommerce catalogs and frequent re-translation.
- Pick Chrome built-in AI if you want zero ongoing cost and your content is not extremely sensitive to nuance. Great for staging environments, internal documentation sites and personal projects. The translation happens in the browser, so nothing leaves your machine.
A hybrid approach works well too. Run Chrome AI or Gemini for the bulk first pass, then re-translate your homepage, key landing pages and top product pages with OpenAI for a quality upgrade where it matters.
Bulk translation tips after running this on 50+ sites
A few things that are not obvious until you have done a few of these:
- Translate pages before posts. Pages are higher-traffic, fewer in number and catch obvious AI mistakes early. You do not want to find out your translation is slightly off after queueing 500 posts.
- Batch posts by category. Run 50 to 100 posts at a time, not 500 at once. Easier to QA, easier to re-run if something looks off and friendlier to your AI provider's rate limits.
- Do WooCommerce in its own pass. Product attributes, variations and checkout strings have specific terminology that benefits from a focused QA pass after translation.
- Do not skip String Translation. Theme labels like "Read more" and "Add to cart", plus plugin strings, all live in WPML's String Translation panel. AutoMLP handles them through the String Translation feature. It is a quick run but it is the one most people forget.
- Cache the responses. If you re-run a batch during testing, AutoMLP's cache saves you from paying twice for the same string.
SEO details that matter for WPML sites
WPML handles hreflang and language URLs cleanly out of the box, but a few SEO details still need your attention regardless of which AI provider you use:
- Translate Yoast or Rank Math meta fields. Page title and meta description need their own translation. AutoMLP handles SEO plugin fields automatically as long as the plugin is detected by WPML.
- Review the translated URL slugs. WPML auto-translates slugs by default. AI sometimes picks awkward translations that look strange in the URL. A quick visual scan of your translated pages saves you from clunky URLs going to production.
- Check translated structured data. If you use FAQ, Article or Product schema, the translated pages should output the schema in the translated language. The String Translation pass handles most of this automatically.
If you also work with Polylang, TranslatePress or Loco Translate sites, the bundle pays for itself on the first project.
Wrapping up
WPML's built-in AI is fine for small sites with light content. The moment you have WooCommerce, multiple languages or any kind of content volume, the credit system stops making financial sense. Bringing your own API key through AutoMLP typically saves 80% to 95% on translation costs and gives you three providers (OpenAI, Gemini, Chrome AI) to match the workload.
If you are still picking a multilingual base plugin, our WPML vs Polylang vs TranslatePress comparison covers which one fits which type of site. For the AI side specifically, the best WPML AI translation addons post covers the third-party landscape in more detail.
