Multi-Language Support

You can run interviews in 18+ languages. The AI speaks the language, the transcription works in that language, and participants get a natural interview experience. This is huge if you're doing global research or want to interview people in their native language.

Supported Languages

We support English, German, Portuguese, Chinese, Japanese, French, Spanish, Hindi, Italian, Korean, Dutch, Polish, Russian, Swedish, Turkish, Arabic, Czech, Filipino, Romanian, and Slovak. Each language has native voice options and transcription support.

How Language Selection Works

Language flows through three levels. When you create an interviewer, you set its language - this is the default for everything. When you create an audience for recruitment, it inherits the interviewer's language by default, but you can override it. Why override? Maybe you've got an English interviewer but want to recruit German-speaking participants. Each actual interview uses its audience's language (or the interviewer's language if there's no audience). This means the same interviewer can run interviews in different languages if you set up multiple audiences.

What Gets Translated

Everything in the interview itself. The AI speaks in the target language, asks questions in that language, and understands responses in that language. The whole conversation happens naturally in whatever language you chose.

We automatically pick the right voice for the language. Different providers (Cartesia, ElevenLabs) have different voices available. Transcription runs in the target language too, so everything gets captured correctly.

If you've set up a pre-interview survey, it gets translated into the target language automatically. Participants see questions in their language, not yours. Email invitations get translated to match the interview language. Standard interface elements (buttons, instructions, etc.) appear in the target language during the interview.

Your admin interface stays in English though. You're creating and managing interviews in English - only the participant experience changes. Research outputs and reports are generated in English for analysis consistency.

Setting Up Multi-Language Interviews

To create a multi-language interviewer: go to your project, create a new interviewer, pick your target language from the dropdown, and configure everything else normally. The language field is right at the top - hard to miss. See Creating Interviewers for full details.

When you create an audience, it inherits the interviewer's language by default. You can change it if needed, and recruitment filters automatically adjust for language + location. Pro tip: Language and country filters work together in recruitment platforms like Prolific. Make sure they match (e.g., German language + Germany/Austria/Switzerland). Learn more in Creating Audiences.

Multi-Language Interview Experience

Here's what happens when someone takes an interview: they click your interview link, if there's a pre-interview survey it appears in the target language, the interview form loads in the target language, the AI interviewer speaks in the target language, transcription captures everything in that language, and the whole flow feels native. No awkward translations. No "powered by Google Translate" vibes.

We automatically pick appropriate voices for each language. The selection includes native-sounding voices for natural conversations, multiple voice options per language, different providers (Cartesia, ElevenLabs) based on availability, and gender and tone variations. You don't have to configure this manually - it just works.

Tips for Multi-Language Research

If you're recruiting through platforms like Prolific, coordinate language with country filters. German speakers are mostly in Germany, Austria, and Switzerland. Spanish speakers could be in Spain or Latin America.

Before running live interviews, do a test. Make sure the AI sounds natural and questions make sense in that language. Different cultures have different conversation norms - what feels friendly in English might feel too casual in German, or too formal in Spanish.

The AI handles language naturally - it's not translating English into German, it's thinking in German. This matters for nuance and idioms. Even though interviews run in other languages, insights get extracted in English. This keeps your analysis consistent across languages.

Common Scenarios

Global Product Research: Create one interviewer in English. Create multiple audiences - one for German speakers, one for Spanish speakers, etc. Same research questions, multiple languages.

Localized User Studies: Create separate interviewers for each language/region. Different focus for different markets.

Mixed-Language Studies: Run some interviews in one language, others in another. Reports aggregate across languages for comparison.

Troubleshooting

Interview sounds robotic: Try a different voice or check if the language is well-supported.

Transcription errors: Some languages transcribe better than others. English, Spanish, German work really well. Less common languages might have occasional issues.

Participant can't understand: Double-check the language is set correctly at the audience level, not just the interviewer level.

Survey not translating: Make sure survey translation is enabled in your interviewer settings. See Pre-Interview Surveys for configuration details.

Technical Details

Behind the scenes: language-specific voice IDs get selected automatically, transcription uses language-aware models, system prompts adapt to language context, and translation services use AI (not simple word-for-word translation). You don't need to know this stuff, but it's why interviews feel natural in each language.