Skip to main content

Review errors (QA flags)

Reviewers flag issues directly on segments as errors, using the industry-standard MQM (Multidimensional Quality Metrics) categories and severity levels. This guide explains how to add them, view them and work through them.

Reviewers add errors by hand, and they're separate from the automatic checks. The editor checks formatting tags and length limits when you save, and underlines misspellings and highlights missing glossary terms as you type, but it can't judge meaning, terminology choices or style. That's what errors are for.

View errors​

  1. In the CAT editor header, click Errors (a bug icon with a count badge if any errors exist).
  2. The All Errors panel opens on the right, listing errors grouped by segment.
  3. Each error shows an MQM category and a severity badge:
    • Critical
    • Major
    • Minor
    • Neutral
  4. Click a segment in the list to jump to it. If nothing has been flagged yet, the panel shows "No errors yet on any segment."
All Errors panel listing three open errors across three segments, each with its severity, category, author and note, and below them the form for adding one to the selected segment #3

A segment with errors also shows a bug icon and count directly on its row, and "Errors (N)" in the row's ⋮ menu.

Add an error​

  1. On the segment's row, open the ⋮ menu and choose Errors. A drawer opens with two tabs, Comments and Errors.

    A segment's ⋮ menu with Comments (2), Errors (1) and the Formatting marks options
  2. Under Add error, pick a Category (Accuracy, Fluency, Grammar, Terminology, Style, Locale Convention, Design or Other) and a Severity.

  3. Add a note if it helps whoever fixes it, then press Enter or click the warning-triangle button.

A segment's drawer on the Errors tab: an open Major terminology error, and the Add error form with Category, Severity and an optional note

The same drawer holds the segment's comments, so a question about a sentence and a flagged error on it sit side by side.

You can also add one without opening the row's menu: select the segment, click Errors in the header, and use the form at the bottom of the All Errors panel. It says which segment the error goes to ("Adding an error to the selected segment #3").

Who hears about it​

The linguist whose wording the error is on gets an email and a notification in the app, with a link to their assignment. That's whoever last changed the segment's translation in their task, or accepted the AI's translation as it was. Your own edits don't count, so if you fix a segment and then add the error, the linguist before you hears about it.

Nobody else on the job is notified. Nobody is notified at all if the latest change was made outside a task (by the AI, or as a correction in the Strings editor), if that linguist is no longer on the task, or if it was made in a client review task.

Errors that Taia's linguists flag are internal: Taia's project manager and the other linguists on the job see them, but your team doesn't.

The email and the in-app notification follow the linguist's own New Message settings in their notification preferences, so each arrives only if they've left that channel switched on.

Resolve a flagged segment​

  1. Click an error (or the segment's bug icon) to jump to that segment.
  2. Edit the target text to fix the issue.
  3. Save the segment with Ctrl+Enter.
  4. Leave a comment on the segment to say it's fixed. Only the person who added an error (or Taia support) can mark it resolved, dismiss it, reopen it or delete it, so the reviewer closes it once they've checked your change.

Submit your work​

There is no submit button inside the editor. Once the flagged issues are resolved, go back to the project page or your assignment page and submit from there. See submit completed work.