Translation Cost Calculator

What actually drives translation cost

Machine and AI translation is priced by volume of text — usually per character or per token — multiplied by the number of target languages. So the two levers are how much content you have and how many languages you want. The calculator above turns pages × words × languages into an estimate across the common providers.

Two very different pricing models

Bring your own key (pay the provider)

You connect an API key from DeepL, Google, or OpenAI and pay them directly for the text they translate. Crucially, this is effectively a one-time cost per piece of content: once a string is translated and cached, serving it again is free. New and changed content is the only ongoing cost.

Per-word SaaS subscription

Cloud services meter your unique word count and bill a monthly subscription, with tiers that step up as you add words and languages. The same 200,000-word, five-language site can sit in a plan that re-bills every month, indefinitely — even if the content never changes. See the difference laid out in Gloty vs. Weglot.

Why the LLM column is so cheap

Smaller LLMs like GPT-4o miniare priced per token at a fraction of classic MT's per-character rate, which is why they often come out cheapest for bulk translation. The trade-off is that quality and latency vary by language and model — DeepL still leads on some European pairs. A good plugin lets you pick per project.

Getting this cost in practice

Gloty is built around the bring-your-own-key model: connect the provider you trust, and every translation is stored in your own WordPress database and reused automatically. Identical text is never paid for twice. Compare approaches against Polylang and WPML.