Is it possible to build a TM from scratch?

Introducing TM Alignment

If you don't have a Translation Memory to share, we can help you build one from scratch based on previously-translated content.

For you to be able to fully benefit from reference documentation and past translations, there are various actions that can be taken depending on the situation to create a new Translation Memory: 

• For document formats that contain identifiers (e.g. .properties files) we will use dedicated alignment tools to automatically segment source and target documents and align (or pair) those segments.
• For document formats that don’t contain identifiers, we'll use AI-based tools to perform the alignment. These tools align segments up to a certain degree of confidence. The lower the confidence, the more segments the tool will be able to automatically align, but with a higher risk of misalignments. Based on your content and resources available for translation, we will recommend or not a manual review of the automatically-aligned pairs and a manual alignment of the segments that the tools were not able to align automatically.

We then import your newly created Translation Memory into the Computer-Assisted Translation (CAT) tool component of our Translation Management System (TMS) and will be used for leveraging for all future projects.

Read more on how we can improve the efficiency and accuracy of your TM.