A few people have asked how Mira actually works in practice, so wanted to write this up.
The full flow from start to finish:
1. Set up your study (5 minutes)
Choose a template Customer Discovery, Concept Testing, UX, Brand Perception, NPS follow-up, and more. Or build your own discussion guide. Mira generates contextual follow-up questions automatically, so you do not need to script every question.
2. Recruit participants (built in)
Access 100M+ participants across 120 countries directly from the platform. Set demographic filters, screener questions, and Mira handles recruitment. No third-party panel needed.
3. Run the AI moderated interview
Participants join via link no app download. Mira moderates the conversation, asks follow-up questions intelligently, and reads facial expressions, voice emotion, and eye gaze in real time during the session. Works on a standard webcam.
4. Get your report (minutes, not days)
Automatic transcript with speaker separation. AI themes, tags, summaries, and key quotes extracted automatically. Emotional signal overlaid on each moment. Full research report generated executive summary, findings, evidence, recommendations.
5. Share and store
AI highlight reels for stakeholders no one watches 40-minute recordings. Everything stored in a searchable research repository. Cross-study intelligence lets you compare findings across multiple projects over time.
The part most people ask about:
The emotional layer runs during the interview not after. So when a participant says "I like it" but their face shows hesitation, Mira catches it and probes deeper in the same conversation. That is the core difference from transcript-only tools.
First study is free this month happy to help anyone set one up. Drop a comment below or book here: https://www.entropik.io/book-dem...
See Mira on Product Hunt: https://www.producthunt.com/post...
What type of research are you running? Happy to walk through how Mira would work for your specific use case.
Mira
Hi Product Hunt 👋
I'm Lava, Founder of Decode by Entropik. We've been building AI that reads human behavior for 9 years and today, we're launching Mira, our AI Moderator.
Here's the problem every researcher knows but nobody talks about: people say one thing and feel another. It's called the Say-Do Gap. Self-reported data is filtered, rationalized, socially edited. Most research tools just accept this. We didn't.
Mira runs the entire research workflow, recruiting, moderating, analyzing, reporting, but uniquely captures what participants say AND feel in real time via facial coding, voice emotion AI, and eye tracking.
When someone says "I love it" but looks confused, Mira notices and probes deeper. Automatically.
Built on 17 patents. 70+ languages. Trusted by Unilever, Nestlé, and 150+ brands.
First study free with code PH20 → entropik.io/platform/ai-moderator
One question for the community: what's the most unreliable part of your current qual research process, and what would it take for you to actually trust AI to run it?
Drop it in the comments. I'll be here all day.
@productrambler What’s one concrete sign that your current qualitative insights are misleading, and what exact safeguards or transparency features would you need to see in an AI moderator before you'd let it probe participants or deliver final insights automatically?
Mira
@swati_paliwal Great question Swati, one we wrestled with a lot when designing the output layer.
A few concrete signs qual insights are misleading:
(1) unanimous positivity — if everyone loves it, your screener may be too narrow or the interview is not challenging enough; (2) verbal/non-verbal mismatch, someone says "I'd buy this" but their face shows confusion or disengagement; (3) vague themes without emotional backing, "ease of use" as a finding with no moments to support it.
On safeguards: Mira shows confidence levels on emotional signals (not just a single call), surfaces the raw emotional timeline so you can verify the signal behind any AI claim, and flags moments of mismatch rather than averaging them out. The AI summary is always a starting point, the original recording, transcript, and emotional data are there for the researcher to check and override.
We are firm that AI should make researcher judgment faster and better supported, not replace it. The final interpretation is always yours.
"Captures what participants say AND feel" is the hard part most tools skip. I work with voice AI daily, and the transcript is the easy 20% — the signal that actually matters (hesitation, tone, the pause before someone says "yeah, it's fine") lives in how they said it, not the words. If Mira genuinely reads that layer at a 100+ panel scale, that's a real moat over "interview + transcript" tools. Question: how do you keep the emotional read from becoming false precision — do you surface confidence levels, or a single sentiment call? Congrats on the launch 🚀
Mira
@david_marko Thank you so much. Since our emotion AI is multi-modal, we use fusion techniques to make sure that false positive are taken care of and yes at raw level we do provide confidence level 😎
Mira
@david_marko Adding to what Sumit said - from a product perspective, the confidence level alongside the emotional call is exactly why we built the output the way we did.
The goal is to give researchers a richer signal to interrogate, not a verdict to accept. The transcript, emotional timeline, and raw clips are always there to verify or challenge what the AI surfaced. It is a tool to improve researchers' judgment, not a replacement for it.
Mira
@ridhwikvinod Ridhwik, this is exactly the right question, and honestly the one we obsessed over the most.
A few things we do to avoid over-reading:
1. Confidence thresholds, not interpolation. If a frame doesn't meet our confidence threshold (lighting, angle, partial occlusion), we drop it entirely rather than fill in the gap. We'd rather have less data than wrong data.
2. Sustained patterns, not moments. A single frown frame means nothing. We look for emotional patterns sustained across 3–5 seconds minimum before flagging them as signal.
3. Triangulation across modalities. Facial expression is one input, we cross-reference with voice tone and eye tracking. Confusion and awkwardness have different voice signatures. That cross-modal agreement is what lifts confidence.
4. Baseline calibration. We establish a neutral baseline in the first 30 seconds of every session so "this person just frowns a lot" doesn't skew the read.
The honest answer is it's not perfect, but neither is a human moderator. The difference is we surface the signal with confidence scores, so researchers can decide what to trust. Happy to walk you through exactly how this works in a live session → https://www.entropik.io/book-demo
The mid-interview probe on the say/feel mismatch is the part that interests me. I run the cheap cousin of this for my own app, a panel of simulated user personas that scores LLM output before a change ships, and the one thing simulation can't give me is exactly that hesitation signal you're reading off real faces. When facial coding and the transcript disagree, which one do your reports trust? I'd want the raw disagreement surfaced, not resolved for me.
Mira
@narek_keshishyan Narek, the fact that you've already built a simulated persona scoring layer tells me you'll get more out of this than most.
To your specific question: the disagreement is surfaced, not resolved. That's intentional and non-negotiable for us.
When facial coding and transcript diverge, someone says "yes, that makes sense" while their face reads confusion and their voice drops engagement — the report shows both signals side by side with timestamps. The researcher sees the verbal response and the emotional trace underneath it. We don't collapse them into a single confidence score or pick a winner. That would defeat the entire point.
What the report flags is the gap itself, the moment where the two signals split. You get the clip, the transcript line, the emotion curve, and a marker that says "these disagreed here." What it means is yours to interpret.
Where we do take a position: we surface which signal is sustained longer. A one-second facial flicker during a considered verbal answer is weighted differently from a 6-second emotional hold that contradicts a polished verbal close. But that weighting metadata is visible, not hidden.
Your simulated personas are great for pre-ship scoring, but the hesitation signal on a real face during a real task is a different class of data. I'd genuinely love to walk you through how the disagreement layer looks in a real study. Can we set up a call? → https://www.entropik.io/book-demo
@mridhu_varshini_ surfaced not resolved is the answer I was hoping for. weighting by how long a signal is sustained makes sense too, a six-second hold is a different animal from a flicker. no call promises mid-launch-week, but I'll poke at the demo materials. appreciate you writing this out properly.
How does the facial coding and eye tracking actually work in a remote setting without making participants uncomfortable or needing specialized hardware?
Mira
@layda138459 Everything runs in the browser — no app download, no plugin, no special hardware. Participants just allow webcam and microphone access when they join, same as any video call. Before the session starts there is a quick calibration check for lighting and camera positioning, which also helps participants get comfortable with the setup before the interview begins. The calibration doubles as a way to normalise the experience — by the time the actual interview starts, most participants have forgotten the camera is doing anything beyond a normal call.
Mira
@layda138459 Everything runs in the browser on a standard laptop or phone, no external hardware and nothing to install. Participants grant camera access upfront, run a short calibration, and the session feels like a normal video call from there.
If someone isn't comfortable on camera, they decline and still complete the study, with emotion read from voice and text instead.
How does the facial coding and eye tracking actually work in practice for remote participants, especially since webcam quality and lighting vary so much across users?
Mira
@ersinyzak1i75 We do not need a very high end webcam to run our models. In fact our model can run on almost all commodity webcams as long as there is good visibility of the face. We however guide the users to adjust face, lights before each session to make sure we capture good data.
Mira
@ersinyzak1i75 This is a great question and we hear this constantly from a lot of our customers during the initial stages. We have designed our system to work with all kinds of webcams, not just the best ones. Before joining, participants go through a quick calibration and environment check. Each video frame receives a confidence score, and we omit any weak frames rather than guessing. We never base our findings on a single video feed, since we combine signals from the entire group. Webcam gaze tracking is less precise than lab infrared, but for questions like whether someone saw a brand element or where their attention went first, it works well. Plus, it lets us study hundreds of real-world settings instead of just a few in a lab. We do not use facial recognition at all; we only measure expressions, not identity.
The facial coding during interviews is genuinely impressive, you can actually see hesitation or excitement that pure transcripts would completely miss. Wish more research tools took emotion seriously like this.
Mira
@mahmutkrc1pj5 Yes, we focus on understanding non verbal cues during interactions and discussions in all of our products. We believe that missing emotion signals might lead to incomplete view of customer's views during interviews and that is why we invested years perfecting multi modal emotion models to provide a comprehensive analysis of customer interviews.
Mira
@mahmutkrc1pj5 Thank you, Mahmut, this is exactly why we built it. The moment before someone chooses their words is often the most honest part of the whole interview. That is the layer that changes decisions.