# QuestionPunk: The survey platform that interviews back.

Canonical page: https://www.questionpunk.com/

Your respondents talk to an AI interviewer that listens, asks the follow-up you would have asked, and turns the conversation into themes and quotes. Text in 130+ languages, voice in 25. Press the mic and be the participant. No sign-up.

**2,300+** researchers have signed up. No sign-up to try · Free to start · No credit card.

Start free: https://app.questionpunk.com?form=signUp

## From one sentence to a full report. In five steps, one sitting.

Describe the study, refine the draft, test it on AI respondents, launch it and read the analysis without leaving QuestionPunk.

1. **Describe.** One plain-English sentence. That’s the whole brief.
2. **Refine.** Edit questions and prompts, pick the AI model, tune follow-ups.
3. **Dry-run.** Pressure-test on 100 AI respondents before you ship.
4. **Launch.** Share a link, embed, or recruit panel participants.
5. **Analyze.** AI insights, themes, quotes, plus full audit trail.

## Type what you want to learn. We’ll design the study.

One plain-English sentence in. A full draft out, question order, types, follow-up logic, and an attention check. You review before a single respondent sees it.

## Real conversations, inside a survey. Text or voice. Same AI.

The AI moderates each interview, asks follow-up questions based on what the respondent just said, and turns the transcript into a report. Run it as a text chat or by voice.

- **Depth control:** Light pulse or deep interview. One dropdown.
- **Attention checks:** Bots and low-effort responses flagged automatically.
- **Adaptive follow-ups:** The AI probes vague answers and digs into interesting ones.
- **130+ languages:** Respondents answer in their language, by voice in 25 of them. You read in yours.
- **Sensitive-info detection:** PII and sensitive info flagged automatically.
- **Voice interviews:** Respondents talk to the AI in their browser. No app download needed.

## Test with 100 AI respondents. Before you recruit a single real one.

Simulate a full cohort before going live. Leading questions, ambiguity, and weak spots are all flagged before a single real participant sees your survey.

## Test your designs. Before you ship the wrong one.

Run pairwise concept tests, AI-moderated discussions and rating scales in one study.

See how concept testing works: https://www.questionpunk.com/features/concept-testing

## A research tool that does the tedious parts. So you can do the thinking part.

Built with researchers who were tired of stitching seven tools together to run one study.

- **Fraud detection:** Catches low-effort, AI-generated, and bot responses automatically. Your data stays clean.
- **Real recruitment:** Target real participants through our panel partners. Demographics, behaviors, interests. Fast.
- **500+ templates:** IRB consent, SLA check-ins, trade show feedback, onboarding studies. Pick one, edit, launch.
- **Multi-language:** Respondents answer in their language. You read reports in yours.
- **Exports your data:** Export to CSV or Excel, or pull responses as JSON through the REST API. No lock-in.

## Need something custom? Talk to a researcher.

If you're a team running regulated research, planning a large rollout, or just want to see the product in action, book a 30-minute call. We'll walk through your use case live.

Book a demo: https://book.questionpunk.com/demo

## Free to try. Fair when you scale.

| Plan | Price, billed monthly | Price, billed yearly | Responses per month | AI credits per month |
|---|---|---|---|---|
| Free | Free | Free | 20 | 3,000 ($3.00 of AI usage) |
| Pro | $50 per month | $500 per year ($41.67 per month) | 5,000 | 25,000 ($25.00 of AI usage) |
| Enterprise | Custom | Custom | Unlimited | Custom AI credits |

Plans, trial terms and AI credits in Markdown: https://www.questionpunk.com/pricing.md

## Frequently asked. Straight answers.

### How does the AI interview work, and how much control do I have?

Three modes: fixed (you set the number of follow-ups), automated (time-based), or the AI decides when it has enough. Choose from 150+ models from Anthropic, OpenAI, Google, xAI, Meta, Mistral, DeepSeek and more. Every prompt is fully customizable and visible to you. No black box.

### Can I run text, voice or video interviews?

All three, even in the same study. Respondents type, talk with a real-time AI voice agent in their browser, or record video. Nobody has to install an app.

### What question types do you support?

50+ types. Standard ones (multiple choice, opinion scale, slider, rating, ranking, matrix, dropdown) plus advanced MRX methods most platforms don't have: MaxDiff (best-worst scaling), conjoint analysis (trade-off testing), monadic concept testing, Van Westendorp (price sensitivity), vignette experiment, list experiment, constant sum, distribution builder, card sorting, heatmap click, and text highlight. Plus AI-native types: AI interview, AI question, AI image generation.

### Can I run my study in multiple languages?

Yes. Auto-translation is built in for 130+ languages, and voice interviews run in 25 of them. Run one study live in many languages at once; respondents pick theirs and the AI follows up in the same one.

### Can I build branching and conditional logic?

Yes. Skip questions, branch on any answer, and route respondents down different paths based on what they said earlier.

### Can it match our brand?

Logo, colors, and fonts. Embed it on your site, share a link, or drop it into a popup.

### How do I get real respondents?

Pick your audience in the app, and we handle recruitment and payment through our panel partners. You can also share a public link or embed the survey on any site.

### What does the AI report give me back?

Themes, quotes, segment breakdowns and an editable narrative. You can trace any insight back to the raw transcripts to check the AI's reading.

### Who owns the data, and what does it cost?

You own your data. Export responses to CSV or Excel, or get them as JSON through the REST API, and delete your account whenever you want. The Free plan includes 20 responses and 3,000 AI credits ($3) a month and needs no credit card. Pro costs $50 a month, or $500 a year billed yearly, and includes 5,000 responses and 25,000 AI credits ($25) a month. Enterprise pricing is custom. You can cancel a paid plan anytime.

## More for AI agents

- [Site index for AI agents](https://www.questionpunk.com/llms.txt)
- [Pricing in Markdown](https://www.questionpunk.com/pricing.md)
- [Help articles in Markdown](https://www.questionpunk.com/support/llms.txt)
- [API and MCP access](https://www.questionpunk.com/support/api-access.md)
- [Comparisons with other survey tools](https://www.questionpunk.com/compare)