Qual AI project screenshot

Qual AI

Live qualitative research platform with AI interpretation and moderated sessions

LiveKitNext.jsNode.jsAI InterpretationRedis

Qual AI is a live qualitative research platform built for pharmaceutical and healthcare studies. It supports the full range of session formats used in the field, including one to one interviews, dyads, triads and focus groups, with separate purpose built interfaces for moderators, respondents, observers and translators so each role only sees what it should.

Sessions are controlled end to end. Role based links, a waiting room, pre call device checks and moderator admission controls decide who gets in, while muting, camera control, participant removal and anonymised display names keep the room manageable once it starts. Moderators work from an interactive discussion guide with a session timer and can save timestamped moments, share their screen, present from a media library and move people into breakout rooms or private debriefing rooms. Frontroom, backroom and private moderator chat channels keep the research team talking without the respondents seeing any of it.

Every session is recorded to the cloud with separate participant tracks and a backup recording, then composed into a single playback that includes speakers, screen shares and presented materials. Live transcription runs with speaker labels and timestamps, and a custom AI interpretation pipeline produces an interpreter audio channel and translated subtitles for observers in real time.

The research tooling covers rating, ranking, multiple choice, open ended and emoji exercises, blind response collection and monadic concept testing, with every response linked back to the participant, the stimulus and the exact recording timestamp. Real time media runs on LiveKit, the backend on Node.js with MongoDB and Redis, and the frontend on Next.js. Regional data storage, customer isolation, reconnection handling and optional face blurring cover the compliance and reliability side.

Key Features

Interviews, dyads, triads and focus groups in one platform
Separate moderator, respondent, observer and translator interfaces
Role based links, waiting room, pre call checks and admission control
Interactive discussion guide, session timer and saved timestamped moments
Breakout rooms, private debriefing rooms and separate frontroom and backroom chat
Cloud recording with separate participant tracks and composed playback
Live transcription with speaker labels and timestamps
Custom AI interpretation pipeline with interpreter audio and translated subtitles
Rating, ranking, multiple choice, open ended and emoji exercises
Blind response collection and monadic concept testing
Regional data storage, customer isolation and optional face blurring

Tech Stack

Next.js

Role specific frontends

LiveKit

Real time audio, video and recording

Node.js

Session, role and exercise backend

MongoDB

Studies, sessions and responses

Redis

Session state and reconnection handling

AI Interpretation

Live translation, interpreter audio and subtitles

Speech to Text

Live transcription with speaker labels

Role

Full Stack Developer

Year

2026

Next Project

AI Video Episode Generator