Context types

Glossary

A glossary defines the exact terminology you expect in your translations and ensures that key terms are used consistently across languages. When you provide a glossary, Lia Translate applies your preferred terms wherever they appear in the text, even when they occur in different grammatical forms.

What format should my glossary be in?

Lia Translate accepts glossaries in all supported formats Supported formats for context, and you can upload bilingual as well as multilingual terminology. The most reliable results come from formats that specify languages and terms very clearly, such as tbx. Other formats like Excel or Word also work, but they may be less precise for lack of structure.

Use a glossary when terminology accuracy is essential—product names, legal terms, technical vocabulary, or branded expressions.

When you upload a multilingual tbx file, Lia will automatically detect the relevant language combination when you use it for translation and extract the relevant terms.

Translation memories (TM)

A translation memory (TM) helps Lia Translate build on work you have already validated. It contains pairs of source and target sentences—typically stored in formats such as tmx, xlsx, or csv—and makes them available whenever the same or similar content appears again. This is invaluable for recurring text, documentation, and any content where long-term consistency is essential.

Lia Translate takes the concept far beyond traditional TM systems. Instead of simply looking for repeated sentences, Lia interprets the tone, style, and phrasing patterns inside your TM and uses them as guidance for new translations. And when a sentence is only partially similar, Lia automatically uplifts the stored translation to match the new context—adjusting grammar, refreshing wording, and aligning it with the current source content. This is how you truly get the maximum value out of your translation memory.

What format should my TM be in?

Lia Translate accepts translation memories in all supported formats, and you can upload bilingual as well as multilingual files. The most structured option is tmx, which clearly defines source and target segment pairs and is widely used across the industry. Other formats like Excel or csv work just as well for guiding consistency, though they may offer less structural precision.

Use a TM whenever consistent phrasing matters—manual updates, evolving product documentation, release notes, or any content that benefits from a unified voice across versions.

Reference and style documents

Reference and style documents help Lia Translate understand how your organization communicates. Instead of defining specific terms or sentence pairs, these documents give Lia examples of your tone, structure, and writing preferences. They act as a guiding framework, allowing your translations to sound more authentic, more aligned with your brand, and more consistent with your existing content.

You can attach any material that reflects how you write: a customer-facing brochure, a tone-of-voice guide, a product overview, internal style rules, or even a well-written sample text. Lia Translate analyses these documents to identify patterns—how formal or informal you sound, how long your sentences typically are, how you structure information, and the overall personality of your content.

Lia Translate then uses these patterns to shape the translation. While the effect depends on the clarity and consistency of the files you provide, even a single well-crafted reference document can noticeably improve the feel and coherence of your output.

What format should my reference or style file be in?

Lia Translate accepts reference and style documents in all supported formats Supported formats for context. You can upload anything from a simple text file to a richly formatted Word document or web-style content. The clearer and more consistent the content is, the easier it is for Lia Translate to detect and apply your stylistic preferences.

Use a reference or style document when tone and personality matter—brand communication, customer-facing texts, editorial content, or any material where the “voice” is as important as the information itself.