Translation Memory
Intelligently reuse approved translations
Significant cost savings
Recurring texts such as standard warnings, regulatory information, or care instructions are automatically retrieved from the translation memory. You never pay twice for the same translation.
Guaranteed consistency
Identical source segments receive identical translations. This is particularly critical for mandatory regulatory statements, warnings, and safety information that must be uniform across all product variants.
Transparent reuse analysis
Before each translation project, manualworks analyzes the expected TM coverage rate. You immediately know what percentage of the document can already be served from the translation memory.
How it works
Build your translation memory
Your translation memory grows automatically with each approved translation project. Additionally, you can import existing TMs from other CAT tools in TMX format to immediately benefit from your existing translation assets.
Automatic matching
For every new document, manualworks automatically matches all segments against the translation memory. Exact matches (100%) are suggested directly, and fuzzy matches (70–99%) are displayed with differences highlighted.
Review and add to inventory
Translators confirm or adjust TM suggestions. After approval, all new and updated segments are automatically written back to the translation memory and are available for future projects.
What Is a Translation Memory?
A translation memory (TM) is a database that stores pairs of source text and approved translations — known as translation segments. When a similar or identical source text appears in a new document, the system automatically suggests the stored translation. For the medical device industry, where many documents contain similar or identical text blocks, this represents an enormous efficiency gain.
Instructions for Use for medical devices typically share a significant proportion of standard text: warnings, regulatory information, symbol definitions according to EN ISO 15223-1, general care and maintenance instructions, and manufacturer contact details. For product families with multiple variants, the overlap between individual Instructions for Use can range from 60 to 80 percent.
Intelligent Matching for Medical Texts
The translation memory in manualworks goes beyond simple segment matching. Our system takes into account the context in which a segment appears: Is the text in a warning, a procedure description, or a technical specification? This contextual information feeds into the ranking of TM suggestions and increases matching accuracy.
Furthermore, the translation memory works seamlessly with the terminology glossary. When a TM hit contains a term that has since been updated in the glossary, the system flags this hit and suggests the update. This keeps your translation memory consistently synchronized with current terminology.
Reuse Across Product Families
A particular advantage for medical device manufacturers: manualworks enables targeted reuse of translations across product families. When you launch a new variant of an existing product, the system automatically analyzes which parts of the existing Instruction for Use can be adopted unchanged and where adjustments are required.
This preliminary analysis gives you a precise estimate of the translation effort and enables reliable budget planning. At the same time, the automatic adoption of approved segments ensures that proven formulations are maintained consistently.
The Value of Your Translation Memory
With each completed project, the value of your translation memory grows. It documents not only translations but also the decisions they embody — which wording was chosen, which term preferred, which variant approved after review. This accumulated translation knowledge is a strategic asset for your organization that makes the translation process increasingly efficient over time.
Frequently Asked Questions
What happens when a source segment changes slightly?+
manualworks uses fuzzy matching technology that suggests matching translations from the translation memory even when there are minor changes in the source text. The similarity level is displayed as a percentage (e.g., 95% match), and the differences from the original segment are color-highlighted. This enables translators to quickly adapt the existing translation rather than translating from scratch.
Can the translation memory be shared between different project teams?+
Yes, manualworks offers flexible sharing options for translation memories. You can create company-wide TMs that all teams can access, or set up project-specific TMs for individual product lines. Access rights can be controlled granularly — for example, external translators can receive read-only access, while internal reviewers can also edit and approve entries.