A survey in two languages is not two surveys.
Run a multilingual survey as one study, with one data table.
However many languages you work in, you prepare one survey. Each language gets its own link, participants answer in their language, and all answers land in one table.
You do not merge two files to compare them: the language of each answer is already in the data.
Do you feel safe in the neighborhood you live in?
هل تشعر بالأمان في الحي الذي تعيش فيه؟
One survey, two links, one data table. In Arabic the screen and layout run right to left.
Thirty languages translated automatically, 150+ written by you.
Draft your survey once. AI translates into major languages. In community or regional languages without automated translation, simply type the questions yourself — it remains one unified study.
Automatic translation
Turkish, English, Arabic, German, French, Russian, Ukrainian, Spanish, Chinese, Hindi and more, with regional variants.
Written by you
Kurdish (Kurmanji and Sorani), Zazaki, Laz, Circassian and many others. No automated translation; you write the questions.
Right-to-left languages
In Arabic, Persian, Hebrew, Urdu, Pashto and Sorani the screen runs right to left, and buttons and arrows follow.
How many languages you can use in one study depends on your plan.
AI drafts the translation; you make the decisions.
The survey title, section titles, questions, help texts, all options and the thank-you screen are translated automatically. What participants see from the system — next, back, error messages — already comes in their language.
You can open any language in the editor; you can edit and refine the translation directly.
Privacy notices are translated separately
Privacy notices and consent texts are not translated with the survey; they go through a separate step and your explicit approval. A legal text cannot be treated like a simple question.
Data protection requirements →Dynamic piping text survives translation
Text that changes with the answer survives translation, but the system does not apply each language's grammar for you; check those sentences in other languages.
Your manual edits stay intact
Your manual edits stay. If you run automatic translation again for that language, the new translation replaces them.
Participants start in their language, and can switch at any time.
Each language has its own link, so nobody is met with a language menu. Still, a participant can switch language mid-survey; the answers given so far are not lost.
The survey picks up seamlessly from the exact same question in the newly selected language.
In the field too
An interviewer switches language on the spot: one interview in English, the next in Arabic, both in the same study.
Live results are multilingual
The reader can switch language, and you can filter results by the language people answered in.
Live survey results →Voice answers in their own language
When a participant speaks instead of typing, the answer is transcribed in the language of the survey. If the language is not set, the system recognises what is spoken.
The language of each answer comes recorded in the data.
The language of each answer comes as a variable in your file, so language is a variable in your data; you can ask directly whether Arabic-speaking and English-speaking participants differ in SPSS.
"Nuvopoll did not start as a software idea. It started inside a study of how two communities see each other. In the first application of the Othering Index, the survey had to run in both Turkish and Arabic — two separate surveys and two tables."
Read our story →Write your survey once, ask in everyone's language.
One survey, a link per language, one data table.