Meet our new AI-powered localization tool

We’ve combined the technological benefits of machine translation with the contextual power of LLMs — and made it work for post-editing bilingual files.

As more and more clients started opting for AI translation, we knew it was time to adapt. That’s why we’ve developed our own AI localization tool — Hera AI, named after our tiniest PM and beloved office assistant.

What it accomplishes:

Hera AI automatically post-edits machine-translated bilingual files using reference materials like translation memories, glossaries, style guides, and any contextual input provided in the prompt tab.

Key benefits:

• Speeds up post-editing while reducing costs
• Maintains high quality through deep contextual understanding by connecting translation memories, glossaries, and style guides
• Processes segments and files in batches to ensure consistent application of contextual input across large volumes of text
• Supports Trados, memoQ, and Phrase file formats, with more options coming soon

First test results from linguists:

• Post-editing speed increased to 600–700 words per hour
• High-quality machine translation, with edit distance up to 92.6%
• Accurate handling of tags and formatting across bilingual files, as Hera AI preserves structural integrity, preventing misplaced tags or formatting issues that often require manual correction
• Approximately 70% consistency in applying terminology, as Hera AI effectively reuses terms from glossaries and TMs

Today, Hera AI is a core part of our AI localization pipeline:

Neural machine translation → LLM post-editing → Native linguist review → Automated QA

This setup gives you fast, scalable, and consistent localization — without compromising on quality.
 

Note that InText respects your choices. We never use AI tools without your explicit approval, keeping you in full control of your projects.
 

We’ll be sharing more soon on how we’ve trained AI to understand translation context based on clients’ reference materials.

Stay tuned!