Fluentia

AI-powered communication coaching through real-time voice roleplay.

Fluentia lets you practice high-stakes conversations — job interviews, client calls, presentations, everyday social situations — with an AI coach that listens, responds, and gives you a detailed performance report after every session.

---

Features

• Scenario-based voice roleplay — choose from curated scenarios across five categories: Interview, Business, Social, Public Speaking, and Everyday
• Live AI voice sessions — speak naturally with an AI agent powered by Vapi + ElevenLabs; the AI adapts its role and tone to the scenario
• Post-session feedback — scored across Clarity, Relevance, and Conciseness with filler-word tracking; report generated by Gemini 2.0 Flash
• Progress dashboard — visualize improvement over time with session history and trend charts
• Coaching preferences — calibrate scoring difficulty (skill level), evaluation style, speaking goals, feedback depth, and feedback tone

---

Screenshots

Scenario Browser	Live Session

Post-Session Feedback	Dashboard

---

Tech Stack

Layer	Technology
Framework	Next.js 16 (App Router) · React 19 · TypeScript
Styling	Tailwind CSS v4 · shadcn/ui · Framer Motion
Auth & DB	Supabase (auth + Postgres)
Voice	Vapi AI · ElevenLabs
AI Feedback	Google Gemini 2.0 Flash
Charts	Recharts
Analytics	Vercel Analytics

---

Getting Started

Prerequisites

• Node.js 18+
• A Supabase project
• A Vapi account with a public key

1. Clone and install

git clone https://github.com/your-username/fluentia-2026.git
cd fluentia-2026
npm install

2. Configure environment variables

Create .env.local in the project root:

NEXT_PUBLIC_SUPABASE_URL=your_supabase_project_url
NEXT_PUBLIC_SUPABASE_PUBLISHABLE_KEY=your_supabase_anon_key
NEXT_PUBLIC_VAPI_PUBLIC_KEY=your_vapi_public_key

3. Run the development server

npm run dev

Open http://localhost:3000.

---

Project Structure

src/
├── app/
│   ├── dashboard/          # All authenticated routes
│   │   ├── practice/       # Scenario browser + live session
│   │   ├── sessions/       # Session history
│   │   ├── feedback/       # Post-session report detail
│   │   ├── profile/        # User profile
│   │   └── preferences/    # Coaching settings
│   ├── authentication/     # Login, register, forgot password
│   └── api/feedback/       # Gemini feedback generation endpoint
├── components/
│   ├── dashboard/          # Dashboard-specific components
│   └── ui/                 # shadcn/ui primitives
├── data/scenarios.ts       # Static scenario definitions
├── hooks/use-vapi-session.ts
└── lib/
    ├── prompts/            # Vapi system prompt builder
    └── voice/              # Vapi client singleton

---

Session Flow

1. Browse scenarios and select one (/dashboard/practice)
2. Review the scenario brief and click Begin Session
3. Speak naturally — the Vapi AI agent plays its role in real time
4. End the session; the transcript is saved to Supabase
5. Gemini 2.0 Flash analyzes the transcript and generates scored feedback
6. Review your report at /dashboard/feedback

---

Available Scripts

npm run dev      # Start development server
npm run build    # Production build
npm run start    # Start production server
npm run lint     # Run ESLint

---

License

MIT