Read with Confidence, Grow with Patience
A calm, playful reading companion designed for children with dyslexia.
WhatDoesDaFoxSay combines speech recognition, micro-expression analysis, and eye-tracking to create a personalized reading experience that helps children practice comprehension — while empowering parents with trust and transparency.
Dyslexia affects 1 in 5 children worldwide, often leading to frustration, anxiety, and loss of self-confidence. Traditional reading tools fail to adapt to how neurodivergent minds process words — especially when compounded by modern distractions like social media and cognitive overload.
WhatDoesDaFoxSay provides a space where learning feels joyful, not judgmental — helping children build patience, rhythm, and reading confidence while giving parents insight into their child’s progress and peace of mind through science-backed transparency.
As a team, we’ve personally experienced or witnessed the struggles of dyslexia — the pauses, the skipped lines, the feeling of falling behind despite trying harder than anyone else.
We asked ourselves:
“What if learning to read could feel more like a conversation — not a test?”
We built WhatDoesDaFoxSay to bridge empathy with technology — turning practice into play and helping children rediscover their voice.
Children read aloud while ElevenLabs listens, transcribes, and identifies tricky sections — analyzing tone, fluency, and stutters in real time.
Using advanced gaze tracking, the app identifies where a reader’s eyes linger or skip — funneling insights from paragraph → sentence → word to pinpoint comprehension gaps.
If the child appears confused, frustrated, or tired, micro-expression analysis triggers adaptive hints or encouragement — making every interaction emotionally intelligent.
A serene space designed for parents — offering articles, scientific studies, and trustworthy stats about dyslexia, ADHD, and child learning.
It’s not a “dark mode.” It’s a peace mode.
Parents can explore the models we use, understand our accuracy, and see the research behind the app’s design.
- Adaptive Learning: Every child receives feedback tailored to their pace and focus level.
- Emotionally Aware: Detects confusion and frustration in real time.
- Parent Trust: Scientific transparency and model explainability build confidence in our platform.
- Data Privacy: All video and audio processing happens securely — none of the data is sold or reused.
- Inclusive Design: Warm, dyslexia-friendly fonts (Atkinson Hyperlegible), pastel farm colors, and gamified UX to reduce cognitive load.
- Early reading intervention for children (Grades 2–4)
- Home or classroom dyslexia practice companion
- Reading fluency tracking for speech therapists and educators
- Parental dashboards for progress tracking and emotional reassurance
- Vanilla HTML/CSS/JS for lightweight performance
- Atkinson Hyperlegible font and farm-themed color palette to reduce reading strain
- Auth0 for secure login via Google, Microsoft, GitHub, Facebook, Passkeys (fingerprint/eye), with ReCaptcha and password-reset via Email/SMS
- FirestoreDB for storing user profiles and reading progress
- ElevenLabs for expressive speech recognition and child-friendly voice synthesis
- Gemini API for adaptive reading comprehension and feedback generation
- OpenFace + MediaPipe Iris for micro-expression and gaze estimation
- Firebase Functions for backend event tracking and analytics
Child reads aloud → audio + webcam input analyzed locally → Gemini summarizes focus + fluency → Firestore stores metrics → Fox Mode displays parental analytics and resources.
- Training real-time micro-expression models to differentiate “confused” vs. “focused”
- Ensuring privacy for minors — running gaze and facial inference entirely on-device
- Creating an emotionally neutral design that feels safe for children yet informative for parents
- Integrating Auth0’s multi-auth methods seamlessly with Firebase and ReCaptcha
- Building a 3-layer funnel (paragraph → sentence → word) pipeline for focus mapping
- Three-Pronged Reading Engine: Real-time fusion of speech, gaze, and emotion for personalized feedback
- Fox Mode: A first-of-its-kind parental space blending transparency, trust, and neuroscience-backed education
- Secure Identity Pipeline: Auth0 + Firestore + ReCaptcha for verified, protected family accounts
- High Accuracy Models: Achieved >90% accuracy in gaze calibration (MediaPipe + Kalman filter) and robust AU detection (OpenFace 3.0)
- Dyslexia is not a limitation — it’s a different pattern of learning that thrives under adaptive design
- Parents need emotional trust as much as children need encouragement
- Accessibility isn’t about simplification — it’s about personalization
- Combining AI with empathy makes education genuinely transformative
- Add voice-emotion modeling for richer feedback during reading aloud
- Integrate fine-tuned stutter detection from the Efficient-Stuttering-Event-Detection model
- Expand Fox Mode into a verified research portal (with direct links to academic studies and impact metrics)
- Build teacher dashboards with class-wide analytics for early intervention
- Deploy mobile-first version with offline reading sessions
Because no child should feel “slow” for learning differently.
Because parents deserve peace of mind, not worry.
And because the fox believes — with patience, every child can learn to read out loud and with confidence.
Sahas, Sai, Kelly, Veer