Don't type - use translation memory and term base

memoQ helps you work faster and stay consistent by suggesting previously translated text (from a translation memory) and known terminology (from a term base).

This page shows you how to get the most out of both and how to keep them clean and efficient.

  • What is a translation memory (TM)?

    A translation memory (TM) stores everything you’ve translated before and reuses it when the same or similar text appears again.

    Example:

    • Your first translation job says: The next stop is Charing Cross → La prochaine station est Charing Cross

    • Your second translation job says: The next stop is Tower Bridge

      So, memoQ shows the previous translation with tracked changes. You only need to update “Charing Cross” to “Tower Bridge.”

    How it works

    1. Translate a segment.

    2. Confirm it with Ctrl + Enter.

    3. memoQ saves it to your project’s TM.

    4. Next time a similar segment appears, memoQ shows a match suggestion in the Translation results pane.

    When you translate in memoQ, the tool looks for similar text in your translation memories (TMs) and LiveDocs corpora. Sometimes the match is perfect, sometimes it’s close, and sometimes it’s only partial. To give you a quick idea of how much editing the match will need, memoQ shows a match rate which is a percentage that tells you how close the stored translation is to your source text.

    Here’s how to read the scores:

    100% Exact match The text is identical.
    101% Context match The text and its context both match.
    102% Double-context match The text, its neighbors, and identifiers all line up. That's the best possible match.
    95–99% Nearly exact Only tiny differences, like punctuation or tags.
    50–94% Fuzzy match The text is similar but needs more editing. High fuzzy (85–95%) means only small changes; low fuzzy (50–74%) means bigger rewrites.

    memoQ may also adjust scores: It can lower them if the source is less reliable, or boost them if it automatically fixes numbers or tags.

    Over time, your translation memories (TMs) can grow large and cluttered. A TM’s size depends on, for example, how many entries it has, or how long they are. If a TM grows too big (around 1 million entries), it can cause resource corruption, slower performance, duplicates, or occasional software crashes.

    How duplicates happen

    Duplicates are usually created when you:

    • Confirm identical translations with different contexts.

    • Use a TM that allows multiple translations for the same source segment and context.

    • Import TMX files that already contain identical source segments.

    • Import the same TMX file more than once.

    • Confirm non-text segments (for example: numbers, dashes, tags).

    How to clean up duplicates in your TM

    1. On Project home, click Translation memories  .

    2. Right-click a TM and choose Edit.

      The Filtering and sorting window opens.

    3. Apply filters and sorting steps to narrow down the entries you want, then click OK.

    4. On the Translation Memory Editor ribbon, click Remove Duplicates .

    5. The Filter for duplicates window appears. Here you decide what counts as a duplicate and what to do with duplicates once they’re found.

    You can try three different option to manage TMs you don’t use at the moment:

    1. Shrink the TM by merging or removing duplicate entries.

    2. Remove old entries

      1. On the Translation Memory Editor ribbon, click Filter/Sort .

        The Filtering and sorting window opens.

      2. Choose a Before date, then export the older entries to a TMX file.

      3. Press Ctrl + Shift + A to select all visible segments.

      4. On the Filtering and sorting ribbon, click Delete  , then Save changes .

      5. Close the Translation Memory Editor.

      6. In the Resource console, import the TMX file into another TM.

        Now you’ll have:

        • A smaller TM for everyday projects.

        • An archive TM with older entries you can use when needed.

    3. Delete a TM

      If you want to delete a TM:

    1. In the Resource console or Project home, right-click the TM and choose Delete.

    2. Confirm by clicking Yes in the confirmation message.

      Delete a TM only if you don't need it anymore. You can't restore it later.

    You can also set up project templates to automatically clean up working memories:

    • After project wrap-up: Confirm and update all segments to the master TM.

    • After project wrap-up: Delete the working TM.

    Regular TM maintenance keeps memoQ running smoothly and prevents crashes or corruption.

    • Number of entries in the TM.

    • Number of duplicates or non-text-only segments.

    • Whether TM folders are stored on an SSD (Solid-State Drive).

    • How many apps are running in the background.

    • Whether TM folders are added as exceptions in your antivirus app.

  • What is a term base (TB)?

    A term base (TB) works like a glossary: it stores terms and their translations. memoQ highlights terms while you translate, helping you stay consistent.

    Adding terms to a TB

    You can add terms while translating:

    • Add with a window

      1. Select a source term in the editor.

      2. Press Ctrl + E.

      3. Type the translation.

      4. Click OK to save.

    • Quick add

      1. Select a source word and its translation.

      2. Press Ctrl + Q.

      3. memoQ saves it instantly — no window needed.

    Using terms while translating

    • Terms appear underlined in the editor.

    • Hover or press Ctrl to see suggestions.

    • Select the right term with arrows and confirm with Enter.

    Term bases can grow large and messy over time. Their size depends on how many entries they contain, languages they cover, and included images.

    Unlike TMs, which use the universal TMX format, TBs can use many formats (CSV, Excel, XML).

    Different CAT tools also handle fields and data types differently. This makes it easier for TBs to grow too large or collect duplicates.

    How duplicates happen

    You can create duplicates when you:

    • Import entries from files

    • Use Quick add term(desktop) or Add term now(web) in the translation editor too often.

    Best practices

    • Keep term bases under 2 GB.

    • Delete TBs you don't need.

    • Add only small images.

    • Clean up duplicates regularly.

    Clean up duplicates in a TB

    1. On Project home, click Term bases .

    2. Right-click a TB and choose Edit.

    3. On the Term Base Editorribbon, click Show Duplicates .

      The Filter for duplicates window opens. Here you can decide what counts as a duplicate and what to do with duplicates once they’re found.

  • How TM and TB work together

    memoQ can merge suggestions:

    • TM provides a full segment suggestion.

    • TB updates specific terms if the previous translation used an outdated word.

    • The suggested segment shows highlighted changes for easy review.

      Always confirm the suggested segment with Ctrl + Enter. Check highlighted terms and any QA warnings.

Pro tips to remember

  • Numbers and tags update automatically in TM suggestions.

  • Always confirm segments with Ctrl + Enter.

  • Always keep your projects and resources in a local folder on your computer.

    Don’t use network drives or cloud storage (OneDrive, Dropbox, etc.). They can disconnect or sync mid-translation and corrupt your files.

Next read

Reuse previous translations with LiveDocs