Skip to main content

QA and LQA in the editor

Reading QA results​

Segments carry quality signals from the QA layers your PM has run:

  • Quality scores — numeric scores from automated scoring; low scores mean "look here first".
  • QA flags — rule-based check failures: tag mismatches, number errors, terminology violations, inconsistencies.

Filter the editor to flagged segments (/qa:fail, /qa:warning) to do a QA-driven pass. Fixing the text and saving clears the stale result on the next QA run. For mechanical issues — punctuation, whitespace, tags — Auto-fix can repair them in bulk.

QA results on segments

Terminology​

When the project has a termbase attached, term matches for the active segment are shown alongside the editor. Terminology QA flags segments whose translation doesn't use the approved term — fix the text or, if the term itself is wrong, comment so the PM can update the termbase.

Logging LQA errors​

On projects with formal LQA review, you annotate errors directly on segments:

  1. Select the problematic part of a translation.
  2. Categorise the error (accuracy, fluency, terminology, style, etc.) and set a severity.
  3. Optionally suggest a correction.

Logged errors roll up into the file's LQA scorecard automatically — you never have to fill in a spreadsheet.

note

The LQA popover opens on top of the segment you're annotating; if you're deep in a long file, the segment row stays highlighted behind it so you don't lose your place.