Enterprise language layer
Your agency, your colleagues and your AI tools write and translate your content. You give each type of content a risk level: the higher the level, the more a person checks. TextUnited follows these levels for every text.
Why now
Until now
An agency or a translation management system (TMS) sent every line to a translator, and often to a reviewer. Quality came from people who read everything. You paid per word for every line and waited for the whole file, regardless of how critical the file was.
Now
Anyone in your company can paste a text into an AI tool and publish the result. The translation arrives at once, but nobody checks it. Nobody compares it with the source or your terms. A text that reads well is not proof that it is correct.
With TextUnited
You set a risk level for each type of content, and TextUnited uses your language data: your approved terms, translations and style guides. TextUnited applies both to every translation it handles. Every approval stays on record in your TextUnited account.
Four risk levels
Your documentation and localization teams assign one of four risk levels to each job. In the standard setup, each risk level leads to one quality level.
TextUnited gives every translated line a quality score, an estimate of how likely an error is. Your team sets the threshold. In one quarter, your teams might send TextUnited texts like these:
Nine languages, published today. A wrong comma harms nobody.
TextUnited translates with your terminology, checks for errors that change the meaning and adds a quality score. You can publish at once when the score is above your threshold. Below it, you decide whether a person checks the text.
Each of these texts needs a different amount of human attention, and TextUnited follows the risk levels that your team sets. The translator or reviewer can be your own linguist, your agency or a TextUnited linguist.
Every term that they approve and every sentence that they correct goes into your language data.
Your language data
Every approved term, every corrected sentence and every style guide is part of your language data.
TextUnited saves it in your account so you can use it in your future translations and when you create content.
T U
Your language data in one place
Your language data stays yours, however you use it.
Headless by design
TextUnited is headless: nobody has to work in a separate translation tool. Your teams and your systems send content from places where they already work. TextUnited applies your risk levels, records who checked each text and routes the work in the background.
Every task stays in one place.
Your CMS, PIM or code repository sends content to TextUnited automatically.
Your n8n workflow starts the job.
Their tools, such as Copilot and Claude, use your approved terms and style guides while they write and translate. They ask for the risk level before they translate through TextUnited.
Systems
Confluence SharePoint Sanity GitHub GitLab
Approved translations go back into these systems.
Formats
PDF IDML DOCX PPTX XLSX XML XLIFF HTML JSON Markdown SCORM
You set a minimum risk level for each content type, for example Critical for every safety instruction.
TextUnited can check every translation before it goes out, and send only the lines with a low score to reviewers.
When an auditor asks who approved a given version of a document and when, the answer is already in your account.
Nobody chases reviewers by email. TextUnited sends each text to the right person and monitors when tasks are done.
We do not use your content to train AI models for anyone else. Your data is stored and processed in the EU.
Test it on your own page
Enter the URL of a live page and the URL of its translation. TextUnited compares the translation with the source. It checks accuracy, missing text, consistent terminology and tone. The report appears at once. You do not need an account or an email address.
The report lists the errors it finds. You decide which errors matter for your content.
For manuals and other documents: book 30 minutes, and we plan a check of one document with you.
I used the files TextUnited provided. Comparing to ChatGPT translations there were significant improvements in translations done in TextUnited.
The translations are fast, precise, and always preserve the original formatting, even in complex documents with tables or technical terminology.
What we do not claim
TextUnited scores every translation. Only the content that needs a person goes to one.
If your old translations contain errors, new translations can repeat them. In Supervised AI translation, TextUnited linguists maintain and improve your language data, from the first import to every later update.
Your risk levels decide which content this is. A person checks every line before you publish it.
Start without switching
Keep your TMS or your agency for the rest. Import your translation memory and termbase as TMX and TBX files, and compare the first jobs with your current process.
Or hand the work to us: in Supervised AI translation, TextUnited takes over your language work for one monthly fee.
