Translation Quality Index (TQI) scores each AI translation in real time, so you can focus review on the translations that need it. Integrated with Transifex AI, TQI uses machine learning to assess translation quality.
TQI is calculated for each AI translation. It evaluates accuracy, grammatical and syntactical quality, and adherence to linguistic style, such as your brand's tone of voice. Use it to focus post-editing, to automate quality assurance, and to deploy content in more languages faster.
TQI calculation
TQI is calculated in real time when you generate translations with AI Tasks or AI Fillups. The score appears next to each translation in the Transifex Editor:
You can also evaluate existing translations with TQI Calculation tasks. Select a batch of translated content and run a quality assessment.
📝 Note:
If you edit a source string and keep its existing translations, or if you edit and save an AI-generated translation, the TQI score for that translation is no longer shown.
TQI scores are not calculated automatically for translations created or edited manually in the Editor. To get a TQI score for these translations, create a TQI Calculation task.
Using TQI
To find AI translations in a specific TQI range, use the filters under More. Choose TQI Greater Than or TQI Less Than and enter a threshold to filter for further review.
TQI Insights and breakdown
You can view a detailed breakdown of how each TQI score is calculated, including recommendations for improving the translation.
Access the detailed analysis
Open the TQI Insights tab in the Editor, next to History, Glossary, and Comments.
Alternatively, when you review a translation generated by Transifex AI, you can:
Click the TQI score shown next to the translation, or
Use the
Ctrl+cmd/alt+Tkeyboard shortcut.
Both options open a window that shows the score for each TQI component.
TQI components
Translation Complexity: Uses agreement between the translations of multiple AI variants as its main signal. It also considers how hard the string is to translate, based on its sentence structure, ambiguity, and terminology density, and it learns from which AI translations reviewers accept and which they edit. Before September 15, 2026, this component was called Translation Consistency.
Structural Integrity: Verifies adherence to formatting, Glossary use, and structural elements like HTML tags to ensure technical correctness. This component is unchanged.
Semantic Quality: Assesses the translation's quality using the MQM framework, evaluating how closely it resembles human-level accuracy. This component is unchanged.
Issues: If the Semantic Quality component identifies any possible issues with the translation, they appear on this tab. Issues are color-coded by severity level (critical, major, and minor), so you can adjust the translation quickly.
If an issue is helpful and informative (or not), click the 👍 or 👎 icon to give immediate feedback. This feedback helps us continuously improve Transifex AI and the algorithm that calculates TQI.
Variants: During AI translation, multiple LLMs generate translations. If a translation isn't quite right, the Variants tab shows a suggested variant and alternative variants that could fit better with what you had in mind.
📝 Note:
TQI analysis is available for translations that have been scored after November 5, 2024.
Variants are available for translations that have been generated after October 30, 2025.
TQI in workflow automation
You can use a TQI threshold to automate parts of your localization workflow. TQI-based automation is available for Transifex AI. It reduces manual steps when translations meet your quality standard. For the other workflow options, see Project workflows.
📝 Note: Since September 15, 2026, TQI scores are stricter. Translations above a threshold need edits less often, and fewer translations may reach the threshold. This is most noticeable at 1.0. Your existing threshold settings carry over.
Available automation options
To turn on TQI-based automation:
Go to your project's Settings.
Open the Workflow tab.
Scroll down to the Quality Assurance section.
Under Review, turn on the TQI-based automation options you want and set a TQI threshold for each target language in your project.
You can configure the following actions based on a TQI threshold:
Automatically review translations: AI-generated translations that meet or exceed the TQI threshold are marked as reviewed and added to Translation Memory (TM), without manual input.
Automatically add unreviewed AI translations to Translation Memory: AI translations that meet the TQI threshold are added to TM even if they have not gone through manual review.
💡 Tip
Looking for more help? Get support from our Transifex Community Forum!
Find answers or post to get help from Transifex Support and our Community.