Skip to main content

Set up a translation project

  1. Projects → Create project. Name it on the Project Name tab.

  2. On Labeling Setup, open Natural Language Processing and pick Machine Translation. The template shows the source text next to a TextArea for the translation.

  3. In the config, change the two Header values to your language pair and rename $english to match your data column (e.g. $source).

  4. On Data Import, upload a CSV/JSON with one row per sentence — the column name must match the $variable in your config. Save.

  5. Optional — Translation Memory and glossary: Project → Settings → Translation (the tab appears on any project whose config has a TextArea). There you can:

    • add TM segments by hand, Import TMX / Export TMX, or Backfill from translations already made in the project;
    • maintain a Glossary of preferred terms, with auto term extraction;
    • toggle Show inline suggestions to annotators and scope the suggestion sources (this project / entire organization / selected projects). Suggestions appear under the translation box while annotators type — click to fill.

    Full card-by-card reference: Translation settings.

Go deeper​