EmotiQ: Your Personal Voice AI Emotional Intelligence Coach

EmotiQ is a voice emotional intelligence coach that uses elevenlabs voice synthesis to play emotional affirmations and guides users in their own voice for micro interventions exercises.

Comment

Inspiration

We were inspired by the growing mental health crisis, particularly among young professionals and students who struggle with stress, anxiety, and emotional awareness. Traditional therapy is expensive and time-consuming, while existing wellness apps often require extensive journaling or lack personalization. We wanted to create something that could provide immediate emotional support through the most natural form of human expression - voice.
 The vision was to build an AI-powered emotional intelligence coach that could understand users' emotional states through voice analysis and provide personalized, evidence-based interventions in real-time. We saw an opportunity to democratize mental wellness by making it accessible, affordable, and engaging through cutting-edge voice AI technology.

What it does

EmotiQ is a comprehensive emotional intelligence coaching app that combines voice emotion analysis with personalized voice affirmations, custom voice affirmation generation and micro-interventions exercises together to create a strong emotional impact for users battling with anxiety, stress, neurodivergent condition (ADHD) and curious users who wants to get detailed insights about their voice

  1. Personal Voice affirmations: Emotional voice affirmations and custom affirmations generations in users voice powered by Elevenlabs instant voice cloning technology. 


  2. Voice Emotion Analysis : it uses Digital signal audio processing and speech recognition to detect emotions from voice recordings, analyzing MFCC features, Average pitch in frequencies, voice energy, Voice quality in decibels (Harmonic to noise ratio) variations of jitter and shimmer, spectral characteristics, and speech patterns. 


  3. Personalized Micro-Interventions : 4 evidence-based exercises (Box Breathing, 5-4-3-2-1 Grounding, Loving Kindness, Progressive Relaxation) with voice-guided instructions for breathing exercises. 


  4. Streak Tracking: Gamified habit formation with weekly (7-day) and monthly (30-day) streak milestones for completed micro-intervention exercises.


  5. Insights: Today/Weekly summaries, emotion trends, voice characteristics analysis, pitch and Energy Trends, Voice Analysis insights, Emotional intensity trends, Emotion Distribution and weekly patterns which can be exported as PDF.

  6. Smart Notifications : Onesignal for notifications and in an app messaging. One signal powers the welcome notification, Emotions analysis detected notifications, weekly (7 )and monthly ( 30 ) completed micro-intervention streak exercises, Emotional goal achievements and Goal reminders notifications, micro intervention in app messages experiences and user behaviour optimal notifications ( 8AM Good morning voice check-ins, streak reminders notifications e.t.c)


  7. Emotional Goal setting with milestones and tracking

  8. Emotional Prompts for self guided reflections.



Target Audience:


  1. Anyone battling with stress and anxiety.
  2. Self-awareness and everyone that wants to set emotional habits.
  3. Wellness newcomers that needs emotional guided support.
  4. Anyone with low self esteem who needs personal affirmations in their voice for emotional impact
  5. Neurodivergent users who benefit from structured, multi-sensory guidance.

How we built it

Technical Architecture:

  1. Frontend: SwiftUI with iOS 17.6+.
  2. Backend: Vercel for backend proxy and security of API keys.
  3. Voice Analysis: Created a dual custom speech recognition and audio processing pipeline with Apple’s Accelerate framework.
  4. Voice AI: ElevenLabs integration for personalized voice affirmations generations and intervention exercises guidance.
  5. Data Persistence: Core Data with lightweight migration support for user data, streaks, and intervention history
  6. Analytics & Push Notifications : One signal for push notifications and user behaviour analyzer.
  7. Monetization: RevenueCat integration for In App purchases ( auto renewing subscriptions and consumables and non consumables purchases). Free trial, $3.99 for premium, $7.99 for pro, $9.99 for 10 affirmations generations and custom affirmations generations. $15.99 for lifetime affirmations generations and custom affirmations generations (users write a scripts, EmotiQ uses the cloned voice to generate audio affirmations)

Challenges we ran into

  • EmotiQ uses the microphone in two different ways so we ran into audio conflict errors. We had to implement an Audio clean up algorithm which is called across views to clean up audio when a user navigates away from an app screen that uses audio. Two instances (The voice check-in tab and the Coaching tab)



- Voice cloning caching for affirmations generations and custom affirmations generations which were burning our Elevenlabs developer credits because playing the audio makes an API request to elevenlabs all the time. 
we implemented a cacheAudioplayerto minimise the API requests.

  • AudioProcessing extraction features: extraction of Audio features such as pitch, energy, spectralcentroid, zerocrossingrate, spectralRolloff, jitter, shimmer, formantfrequencies, harmonicToNoiseRatio, voiceOnsetTime and mfcc coefficients from audio recordings for emotion analysis was a major challenge because these values were used in the calculations of Digital signal processing algorithms.

  • Fitting audio extracted features into charts for visualisation was also technical challenging. 
Voice synchronisation: Matching breathing visualisations with Elevenlabs cloned voice for breathing voice guidance was also technical challenging.

  • Persistence of emotional Data, voice check-ins and streaks for micro-interventions exercise was an hassle to get right.


  • We ran into Onesignal UnAuthorisation headers a lot when we were migrating from local development to remote backend proxy.

Accomplishments that we're proud of

  1. We are proud to have built a creative and innovative production voice emotion analysis system.

  2. we are proud of the Implementation of streak counting for all 4 micro-intervention exercises and passing the results as push notifications and in app messaging upon completion. 

the seamless voice-guided intervention with synchronized breathing visualizations and haptic feedback is feature we are really really proud because we use the user's cloned voice in a creative way to give them intuition.

  3. We are proud to have use Elevenlabs instant voice cloning to provide voice guided quick reliefs to users such box breathing and 4-7-8 breathing exercises. 



  4. We are proud of developing of an insight feature which shows users the analysis of their voice features which can be exported as PDF.

  5. we created an easy to use and intuitive, accessible mental wellness tool that requires minimal user effort.

What we learned

  1. Voice emotion analysis requires robust validation of audio features to prevent false positives
  2. Core Data persistence is crucial for habit-forming apps, users expect their progress to be maintained
 Micro-interventions (2-5 minutes) are more effective than longer sessions for busy users.
  3. we learnt about Creating multipart form data for voice upload to use Elevenlabs Service.
  4. We learnt how to make the most of OneSignal SDK and API and ElevenLabs API.

What's next for EmotiQ: Your Personal Voice AI Emotional Intelligence Coach

  1. Internationalization and localization: We want to expand EmotiQ to support multiple languages. 


  2. Market expansion: create the android version of EmotiQ.
  3. 

Launch survey campaign to determine the features users interact with or features users want to see in the app.
  4. launch advertising campaign for EmotiQ.

Built With

Share this project:

Updates