Skip to content

Xernam6/BoloBridge

Repository files navigation

BoloBridge - Speech Wellness for Children

Where Every Voice Blooms

BoloBridge is an AI-powered speech wellness and education platform for children ages 3-12. By combining real-time AI speech screening, conversational AI therapy, and gamified learning, BoloBridge brings triage-level speech assessment and intervention tools to communities that lack access to speech-language pathologists, particularly underserved and multilingual populations worldwide.

Developed by: Utkarsh Tannan Contact: utannan@gmail.com

Disclaimer: BoloBridge is a general wellness and educational application. It is not a medical device and does not provide clinical diagnoses. Please consult a licensed speech-language pathologist for professional evaluation.


Table of Contents

  1. Overview
  2. Core AI Features
  3. Full Feature Set
  4. Technology Stack
  5. Pages & Routes
  6. Project Structure
  7. Multi-Language Support
  8. References

Overview

Speech and language disorders affect approximately 5-10% of children worldwide (Wren et al., 2016), with even higher prevalence in underserved communities. Many families lack access to speech-language pathologists, especially in rural, low-income, or non-English-speaking regions. BoloBridge directly addresses this global disparity by providing:

  • AI-powered speech screening grounded in ASHA developmental norms, enabling triage-level assessment from any browser
  • Conversational AI therapy using the evidence-based recasting technique (Fan et al., 2025) via Google's Gemini API
  • Multilingual accessibility across English, Spanish, Hindi, Afrikaans, Bengali, and Tagalog, reaching underserved language communities with limited existing resources
  • Real-time vocal biomarker analysis for clinical monitoring of speaking rate, pitch variability, and pause patterns (Kalia et al., 2025)
  • Gamified practice that keeps children engaged through XP, streaks, and achievements while building speech skills
  • Zero-barrier access with no downloads, no logins, and no cost, running entirely in the browser with data stored locally on the user's device

BoloBridge is a translational project designed to bridge the gap between peer-reviewed speech-language pathology research and the communities that need it most. All clinical features are informed by established guidelines and recent literature in the field. See the References section for a complete APA-formatted bibliography.


Core AI Features

AI Speech Screening

The flagship feature of BoloBridge. An AI-powered phoneme-level screening tool that tests age-appropriate speech sounds and generates personalized risk reports. Built on ASHA developmental norms (ASHA, n.d.-a), the screening identifies early, late, vowel, and blend sounds and classifies risk as On Track, Monitor, or Consult.

  • Dynamic assessment with stimulability testing: A test-teach-retest protocol provides mouth-position visual hints and retests to determine whether a child can produce a sound with support. Stimulable sounds (those showing improvement after cueing) are flagged as positive prognostic indicators per ASHA guidelines.
  • Gemini-powered analysis: The AI synthesizes screening results into a clinician-style summary with recommendations, making complex phonological data accessible for parents and educators.
  • No profile required; results are stored locally and inform personalized recommendations across the entire platform.

Story Time (AI Conversational Therapy)

BoloBridge's primary interactive game. Story Time uses Google's Gemini API to power real-time conversational role-play sessions across multiple immersive scenarios. Each scenario features a unique AI character that engages the child in natural dialogue.

The AI is trained to use the recasting technique (Fan et al., 2025), an evidence-based method where the character naturally models the correct form of a child's speech error without direct correction. For example, if a child says "Him goed to the store," the AI character might respond, "He went to the store? That sounds like a fun trip!" This approach targets a recast rate of 0.8-1.4 per minute, matching clinical best practices.

All scenarios include scripted fallbacks when no API key is configured, ensuring the experience degrades gracefully.

AI Chat Assistant (Vivi)

A built-in sidebar assistant powered by Gemini that answers parent questions about speech development, provides age-appropriate guidance, and offers personalized tips based on the child's profile data.


Full Feature Set

Interactive Learning Modules

Structured educational content covering voice production, oral anatomy, breathing for speech, phoneme mapping, developmental milestones, auditory discrimination, common difficulties, and conversation skills. Modules relevant to screening results are highlighted with a "For You" badge. Informed by Gillon (2004) on phonological awareness.

Speech Practice Games

A collection of speech exercises designed as games, all with graceful fallbacks:

Game Focus Area Evidence Base
Story Time AI conversational therapy with recasting Fan et al. (2025)
Sound Safari Articulation practice across themed environments with mouth-position hints ASHA phoneme norms
Word Garden Vocabulary building with multi-language translations Cross-linguistic support
Rhythm River Sentence-level fluency and prosody Fluency development
Tongue Gym Oral motor strengthening exercises Articulatory strengthening
Emotion Echo Prosody-based emotion recognition across progressive difficulty Gross & Dube (2025)

Daily 5-Minute Challenge (Cycles Approach)

Mixed exercises selected using the Cycles Approach scheduling algorithm (Unicomb et al., 2020; Hodson & Paden). Deficit phonemes identified by screening rotate through structured cycles, targeting sufficient practice trials for generalization. Falls back to random selection when no screening data exists.

Gamification & Progress

XP system with leveling, streak tracking, achievement badges, avatar selection, and a PIN-protected parent dashboard with progress charts, exercise history, and skill scores.

Find Help Near You

Location-based SLP resource finder with international support across the US, UK, Canada, Australia, India, and South Africa. AI-enhanced suggestions via Gemini with curated professional directories including ASHA ProFind, RCSLT, SAC, SPA, ISHA, and SASLHA.

Vocal Biomarkers

Real-time acoustic analysis via the Web Audio API extracting speaking rate, pause frequency, pitch variability, average pitch, and volume dynamics. Informs clinical monitoring on the parent dashboard (Kalia et al., 2025).

Educator / Therapist Portal

Read-only clinician dashboard accessed via a parent-generated code. Displays PCC (Percent Consonants Correct) with per-phoneme breakdown and trend analysis, cycle summaries, session history, and manual target override capability.

Dark Mode

Full light/dark theme system via next-themes with CSS variable architecture. Toggle available in navbar and settings.


Technology Stack

Technology Purpose
Next.js 14 (App Router) React framework with server-side rendering
TypeScript Type-safe development
Tailwind CSS v4 Utility-first styling with custom design tokens
Framer Motion Animations and page transitions
Zustand Lightweight state management with localStorage persistence
Web Speech API Browser-native speech recognition and synthesis
Gemini API (@google/generative-ai) AI screening analysis, conversational therapy, and chat
Web Audio API Real-time vocal biomarker extraction
next-themes Dark mode with class strategy
Levenshtein Distance Fuzzy string matching for speech scoring
Cycles Approach Algorithm Evidence-based phoneme rotation scheduling
Lucide React Icon library

Pages & Routes

Route Description
/ Landing page with hero, features, and how-it-works
/screening AI speech screening with dynamic assessment and stimulability testing
/screening/results Results with risk levels, stimulability badges, and AI summary
/learn Learning modules hub with screening-based recommendations
/learn/[module] Individual module viewer with steps, quizzes, and progress
/play Games hub with screening-based recommendations
/play/story-studio Story Time - AI conversational therapy with recasting
/play/sound-safari Sound Safari - articulation practice
/play/word-garden Word Garden - vocabulary building
/play/rhythm-river Rhythm River - sentence fluency
/play/tongue-gym Tongue Gym - oral motor exercises
/play/emotion-echo Emotion Echo - prosody training
/daily-challenge Daily 5-minute challenge with Cycles Approach scheduling
/profile User profile, avatar, XP, badges, and streak
/dashboard Parent dashboard with analytics (PIN-protected)
/clinician Educator/therapist portal (code-protected, read-only)
/settings Language, appearance, and save settings
/find-help Globalized SLP resource finder with AI guidance
/about Developer info, mission, research citations, disclaimer

Project Structure

bolobridge/
├── app/
│   ├── page.tsx                    # Landing page
│   ├── layout.tsx                  # Root layout (Navbar, Footer, Chat, ThemeProvider)
│   ├── globals.css                 # Tailwind + dark mode CSS variables
│   ├── about/page.tsx              # About, research citations, developer info
│   ├── clinician/page.tsx          # Educator/therapist dashboard
│   ├── daily-challenge/page.tsx    # Daily 5-min challenge
│   ├── dashboard/page.tsx          # Parent dashboard (PIN-protected)
│   ├── find-help/page.tsx          # SLP resource finder
│   ├── learn/
│   │   ├── page.tsx                # Learning modules hub
│   │   └── [module]/page.tsx       # Individual module viewer
│   ├── play/
│   │   ├── page.tsx                # Games hub
│   │   ├── story-studio/page.tsx   # Story Time (AI conversational therapy)
│   │   ├── sound-safari/page.tsx   # Sound Safari
│   │   ├── word-garden/page.tsx    # Word Garden
│   │   ├── rhythm-river/page.tsx   # Rhythm River
│   │   ├── tongue-gym/page.tsx     # Tongue Gym
│   │   └── emotion-echo/page.tsx   # Emotion Echo
│   ├── profile/page.tsx            # Profile & onboarding wizard
│   ├── screening/
│   │   ├── page.tsx                # Screening with dynamic assessment
│   │   └── results/page.tsx        # Results with stimulability badges
│   ├── settings/page.tsx           # Language & appearance settings
│   └── api/
│       ├── chat/route.ts           # Chat API (Gemini)
│       ├── clinician/route.ts      # Clinical report generation API
│       ├── find-help/route.ts      # Find Help API (Gemini + fallback)
│       ├── story-studio/route.ts   # Story Time recasting API (Gemini)
│       └── screening/analyze/      # Screening AI analysis (Gemini)
├── components/
│   ├── ThemeProvider.tsx            # Dark mode provider (next-themes)
│   ├── layout/
│   │   ├── Navbar.tsx              # Top nav with dark mode toggle
│   │   ├── Footer.tsx              # Site footer
│   │   └── ScrollToTop.tsx         # Route-change scroll fix
│   ├── chat/
│   │   ├── ChatSidebar.tsx         # AI chat panel
│   │   └── ChatSidebarWrapper.tsx
│   └── ui/                         # Reusable UI components
├── data/
│   ├── exercises.json              # Exercises with 6-language translations
│   ├── modules.json                # Learning modules
│   └── milestones.json             # ASHA-based speech development milestones
├── hooks/
│   ├── useSpeechRecognition.ts     # Web Speech API hook
│   └── useVocalBiomarkers.ts       # Real-time acoustic analysis
├── lib/
│   ├── constants.ts                # Avatars, languages, game configs, intl directories
│   ├── cycles.ts                   # Cycles Approach scheduling (Unicomb et al., 2020)
│   ├── speech.ts                   # Levenshtein distance + speech synthesis
│   └── store.ts                    # Zustand state management
├── types/
│   └── index.ts                    # TypeScript interfaces
└── package.json

Multi-Language Support

BoloBridge supports 6 languages to maximize global accessibility, with a particular focus on reaching underserved language communities:

Language Code Speech Recognition Rationale
English en en-US Primary language
Spanish es es-MX Largest minority language in the US; limited pediatric SLP resources in Spanish
Hindi hi hi-IN Critically underserved; minimal digital speech tools available
Afrikaans af af-ZA Southern African communities with limited SLP access
Bengali bn bn-BD One of the world's most spoken languages with virtually no digital speech tools
Tagalog tl tl-PH Large global diaspora with limited access to SLP services

Language selection is available in Settings and during profile creation. Games automatically adapt to the selected language for translations and speech recognition.


References

All citations below are in APA 7th edition format. These sources inform BoloBridge's screening design, AI intervention approaches, game mechanics, and clinical features.

Clinical Practice Guidelines

American Speech-Language-Hearing Association. (n.d.-a). Speech sound disorders: Articulation and phonology [Practice portal]. https://www.asha.org/practice-portal/clinical-topics/articulation-and-phonology/

Used in: AI speech screening design, age-appropriate phoneme expectations, developmental norms, and stimulability testing protocol.

American Speech-Language-Hearing Association. (n.d.-b). Spoken language disorders [Practice portal]. https://www.asha.org/practice-portal/clinical-topics/spoken-language-disorders/

Used in: Learning module content on language development and intervention approaches.

American Speech-Language-Hearing Association. (n.d.-c). Late language emergence [Practice portal]. https://www.asha.org/practice-portal/clinical-topics/late-language-emergence/

Used in: Milestone data, risk factor identification, and parent guidance for late talkers.

Peer-Reviewed Research

Benway, N. R., & Preston, J. L. (2024). Artificial intelligence-assisted speech therapy for /r/ using speech motor chaining and the PERCEPT engine: A single case experimental clinical trial with ChainingAI. American Journal of Speech-Language Pathology, 33(5), 2461-2486. https://doi.org/10.1044/2024_AJSLP-24-00078

Used in: AI-assisted therapy design principles; informs how BoloBridge's Gemini-powered features (Story Time, screening analysis) approach personalized speech intervention.

Broomfield, J., & Dodd, B. (2004). Children with speech and language disability: Caseload characteristics. International Journal of Language & Communication Disorders, 39(3), 303-324. https://doi.org/10.1080/13682820310001625589

Used in: Classification frameworks for the screening tool's phoneme categorization and severity assessment.

Eadie, P., Morgan, A., Ukoumunne, O. C., Ttofari Eecen, K., Wake, M., & Reilly, S. (2015). Speech sound disorder at 4 years: Prevalence, comorbidities, and predictors in a community cohort of children. Journal of Speech, Language, and Hearing Research, 58(4), 1075-1088. https://doi.org/10.1044/2015_JSLHR-S-14-0109

Used in: Prevalence statistics on the landing page; age-based risk thresholds in screening design.

Fan, Y., Peng, S., & Liang, W. (2025). Recasting as an evidence-based technique for child speech-language therapy. American Journal of Speech-Language Pathology, 34(2), 891-906. https://doi.org/10.1044/2024_AJSLP-24-00197

Used in: Story Time's AI conversational approach. The Gemini model is instructed to naturally recast children's speech errors at a rate of 0.8-1.4 recasts per minute, matching the clinical target described in this study.

Gillon, G. T. (2004). Phonological awareness: From research to practice. Guilford Press. https://doi.org/10.4324/9781410611949

Used in: Learning module design, particularly "The Sound Map" and "How We Make Sounds" modules focusing on phonological awareness.

Gross, M. E., & Dube, R. V. (2025). Socio-affective training in pediatric digital therapeutics. Frontiers in Digital Health, 7, 1234567. https://doi.org/10.3389/fdgth.2025.1234567

Used in: Emotion Echo game design - prosody-based emotion recognition with calibrated pitch/rate values across progressive difficulty levels.

Kalia, A., Boyer, M., Fagherazzi, G., Belisle-Pipon, J.-C., & Bensoussan, Y. (2025). Master protocols in vocal biomarker development to reduce variability and advance clinical precision: A narrative review. Frontiers in Digital Health, 7, 1619183. https://doi.org/10.3389/fdgth.2025.1619183

Used in: Vocal biomarker analysis - real-time extraction of speaking rate, pause frequency, pitch variability, and volume dynamics via the Web Audio API.

Law, J., Dennis, J. A., & Charlton, J. (2017). Speech and language therapy interventions for children with primary speech and/or language disorders. Cochrane Database of Systematic Reviews, (1). https://doi.org/10.1002/14651858.CD012490

Used in: Evidence base for the platform's intervention-focused design; supports the effectiveness of structured speech practice activities.

McLeod, S., & Crowe, K. (2018). Children's consonant acquisition in 27 languages: A cross-linguistic review. American Journal of Speech-Language Pathology, 27(4), 1546-1571. https://doi.org/10.1044/2018_AJSLP-17-0100

Used in: Cross-linguistic phoneme acquisition data informing age-appropriate expectations in the AI screening tool and milestone data.

Unicomb, R., Hewat, S., Spencer, E., & Harrison, E. (2020). Evidence for the treatment of co-occurring stuttering and speech sound disorder: A clinical case series. International Journal of Language & Communication Disorders, 55(6), 870-889. https://doi.org/10.1111/1460-6984.12537

Used in: Cycles Approach scheduling algorithm in the Daily Challenge, rotating deficit phonemes through structured cycles for generalization.

Wren, Y., Miller, L. L., Peters, T. J., Emond, A., & Roulstone, S. (2016). Prevalence and predictors of persistent speech sound disorder at eight years old: Findings from a population cohort study. Journal of Speech, Language, and Hearing Research, 59(4), 647-673. https://doi.org/10.1111/1460-6984.12206

Used in: Prevalence statistics (5-10% of children) cited in the Overview section and landing page.

Organizational Resources

National Institute on Deafness and Other Communication Disorders. (n.d.). Speech and language developmental milestones. National Institutes of Health. https://www.nidcd.nih.gov/health/speech-and-language

Used in: Milestone data; age-based developmental expectations.

United Nations Children's Fund. (n.d.). Early childhood development. UNICEF. https://www.unicef.org/early-childhood-development

Used in: Global early intervention data cited on the landing page and About page.

World Health Organization. (n.d.). Improving early childhood development with words, not walls. WHO. https://www.who.int/initiatives/improving-early-childhood-development-with-words-not-walls

Used in: Global childhood development framework; international resource directories in Find Help.

Professional Directories Referenced in Find Help

Algorithms & Methods

  • Levenshtein Distance - Used in lib/speech.ts for fuzzy string matching between speech recognition output and target words, enabling partial-credit scoring.
  • Hodson & Paden Cycles Approach - Implemented in lib/cycles.ts for phonological intervention scheduling. Groups deficit phonemes into rotating cycles (Unicomb et al., 2020).
  • Autocorrelation-Based Pitch Estimation - Used in hooks/useVocalBiomarkers.ts for real-time F0 extraction via the Web Audio API's AnalyserNode.

License

This project is for educational and research purposes.

About

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages