Multify Blog

Automatic AI Translation and Multilingual Functionality on Tilda: How Multify Works

Using modern AI models — such as GPT, DeepSeek, Mistral — has taken website translation automation to a completely new level. This is especially relevant for Tilda websites, where multilingual functionality is not supported out of the box, and automatic translations from Google, Yandex, etc. require manual adjustments, which is a very labor-intensive and error-prone process, and also do not provide the desired results for SEO optimization of the site, since they translate the site after it has loaded in the browser.
The Multify translates text not just with single requests to AI models, but uses a whole system - taking into account context and graph architecture, which helps maintain the coherence of the text throughout the site. Here's how it works.

🔄 Why context matters

When translating text on a site, especially a small fragment (e.g. a menu item, button or footer line), an isolated approach gives inaccurate results. The model may not understand what the phrase refers to, how it is consistent with other elements, and choose the wrong translation.
To avoid this, Multify passes to the model not only the text being translated, but also its contextservice:
[ text before ]
[ text being translated ]
[ text after ]
This approach helps the model to “see” a fragment not as a disjointed phrase, but as part of a coherent whole.

📍Example

On this site, from the button "more about" the model takes into account the context of the upper and lower blocks when translating:
In the code below, the considered context around the button is highlighted "more about"service:

➰ Graph structure: how context is formed

To determine the environment of fragments even more precisely, Multify breaks down the entire document (web page) into blocks and forms a bidirectional graph from them. This means that:
  • Each text element knows which blocks are nearby.
  • If a new fragment is added in the middle of a page, its “neighborhood” can be automatically determined.
  • The model receives not only the fragment itself, but also the logically related blocks — even if they were translated earlier.
This approach helps maintain semantic integrity, as well as grammatical consistency — for example, correct cases, tenses, and stylistic uniformity.

💯 Why does this work better?

Context in translation is the key to quality. This is especially noticeable in:
  • Complex names and technical terms
  • Short phrases without verbs (e.g., «For home», «To warehouse»)
  • Repeating elements that depend on their surroundings
Without understanding the context, LLM can generate a "formally correct" but unnatural or incorrect translation. Thanks to the graph and correct transmission of the environment, Multify avoids these errors.

🔝 What's the difference for SEO?

In addition to the quality and accuracy of translation using AI models, there is also a significant difference in the technical implementation of a multilingual website, which greatly affects search engine rankings. The thing is, translating a website using Google Translate or Yandex Translate, which is done after the page is loaded in the browser, such translations are performed on the client-side and do not contribute to SEO optimization.

Search engines do not index translated content, created using automatic tools without human editing, as it is considered automatically generated content.
"Google не индексирует переведённый контент, созданный с помощью Google Translate, что ограничивает видимость вашего сайта на международных рынках."
→ Source: Auris AI [translation mine]
Thus, for effective SEO-optimization of a multilingual site it is recommended to use server-side solutions, which provide search engines with access to translated content.

🚀 Claude — best translation quality

In Multify currently uses Claude 3.5 Haiku — this LLM shows the best translation quality for languages of CIS countries: Kazakh, Uzbek, Ukrainian, Romanian, Azerbaijani, Kyrgyz, Armenian, Tajik, Belarusian, Turkmen.
Despite Claude's high cost compared to other AI models, thanks to its proprietary architecture and the use of unlimited servers, Multify offers:
💸 Competitive pricing
🥇 The highest quality among Tilda solutions
🌐 Support for complex languages and regional variations

🎯 Summary: how it all works together

  1. The page is broken down into logical chunks
  2. Each chunk gets its own context — text before and after
  3. A graph is formed, so that when updates are made, the surrounding context can be quickly identified
  4. The model receives the necessary context and produces a coherent, accurate translation
As a result, you get:
✅ A multilingual site without duplicates
✅ A translation that reads naturally
✅ Improved SEO tags and meta tags
✅ Flexibility and Scalability without manual routine
2025-05-16 15:31 Multify Functionality