Auto QA is available for both the CAT and TMS modules.
Auto QA allows you to set up rules against which translations will be automatically checked. This article will guide you on how to configure Auto QA to boost your LQA speed and quality while working in the CAT editor and TMS Jobs.

Create new Auto QA rule setting
- In Gridly homepage, select Localization settings.
- Switch to the Auto QA tab.
- Click Create New Setting.

To assign Auto QA rules setting to a project in CMS module:
- Click Select a project to assign.
- Select projects from the list.

To assign Auto QA rules setting to a project in TMS module, refer to this guide.
Project assignment here can be overruled by the settings in CMS Database or TMS Category.
Customize Auto QA rule setting
- Click Customize.
- Here you will see 6 tabs representing different rule groups. Go through each tab and select the rules you want to include.
- Segment and terms: Checks for terminology consistency, translation completeness, and content-related issues within individual segments.
- Consistency: Ensures translation consistency by checking for uniform terminology usage and consistent translations of identical source segments.
- Inline tags: Validates proper handling and maintenance of inline tags.
- Spelling and grammar: Checks for spelling errors and grammatical mistakes in your translations.
- Gridly AI: Use Gridly AI to identify issues in your translations.
- Severity settings: Configure the severity level for each Auto QA rule. You can set each rule to one of three severity levels, including Error, Warning, and Information. Remember to Save your settings here.

Make sure to select at least one rule for your Auto QA rule setting. If none is selected, you won't be able to run Auto QA checks.
Perform Auto QA check in the CAT editor
- Open the CAT editor.
- Click the
button on the right-hand side toolbar to open QA check panel.
- Click the
button and make sure Auto QA is selected.
- Click Run QA check.

Learn more about QA status in the CAT editor and how to interpret QA results.
Auto QA report in TMS Jobs
Auto QA report in TMS Jobs allows you to have an overview of Auto QA runs performed in the Job's Tasks. This allows you to have the most detailed insights on translation quality of TMS Jobs.
Auto QA reports are only available for jobs whose category type is set to Gridly CAT.
In a TMS Job, click QA Report to view Auto QA report. The Auto QA report has 2 tabs: Overview and Detailed.

Overview tab
This tab displays the latest Auto QA results for each Task in the Job, including:
- Language pair
- User who performed the check
- Number of errors, warnings, and information items found
In this tab, Tasks is grouped by Workflow step, which is determined by the Job's Workflow template. You can filter which Tasks to view using the Workflow step dropdown menu. Click on the language pair to open the Task in Gridly CAT editor.

Detailed tab
This tab displays all Auto QA runs performed for Tasks in the Job.

First, filter which Task whose Auto QA report you want to see using the Language pair and Workflow step dropdown menus. Click Open task to open it in the CAT editor.

You will see the report of all Auto QA runs performed in the Task, including:
- When each Auto QA check was executed.
- Which user performed the Auto QA check.
- Number of errors, warnings, and information items found.
For easy comparison, Gridly highlights the Latest QA result, the Previous run, and the First run.

The report also indicate the changes in number of issues between runs:
- Red numbers: Increase comparing to the previous run
- Green numbers: Decrease comparing. to the previous run

To view the detailed result of each run, click on its run time. To learn more about each error code, refer to this article: Auto QA errors reference guide.

QA Report status
You can quickly determine if a TMS Job still has unresolved Auto QA issues using the QA report status:
If the status is orange, the Job still has unresolved Auto QA issues.

When delivering a TMS Job with unresolved Auto QA issues, you will encounter this warning message.

If the status is green, the Job doesn't have any unresolved Auto QA issues and can be delivered without any warnings.
