Unmute

Because everyone deserves to be heard.

Unmute is an AI-powered communication assistant designed for autistic and nonverbal individuals. Unlike traditional AAC tools, Unmute adapts dynamically—predicting user intent, providing natural voice synthesis, and learning user-specific expression patterns.

---

✨ Features

• Adaptive AI Communication – Uses GPT-OSS to predict intent and anticipate relevant phrases that fit the user’s style.
• Expressive Speech – NVIDIA-powered synthesis delivers emotional, natural-sounding voice output.
• Personal Growth Engine – Learns evolving vocabulary and a user’s communication style over time.
• Offline-Ready PWA – Service worker, manifest, and offline fallback keep the experience responsive even with limited connectivity.
• Open Source – Apache-2.0 licensed and community-driven.

---

🧭 Project Structure

This repository contains the Next.js 14 PWA in the root directory:

app/                # App Router pages, routes, and metadata
components/         # Reusable UI building blocks
lib/                # Shared utilities and NVIDIA API helpers
public/             # Static assets, manifest, service worker, offline page
next.config.mjs     # Next.js configuration
postcss.config.mjs  # Tailwind/PostCSS configuration
tailwind.config.ts  # Tailwind design system

The .github/workflows/ directory now only hosts GitHub Actions workflow definitions.

---

🚀 Quick start

git clone https://github.com/Mjs-Apps/unmute.git
cd unmute
npm install
cp .env.example .env
# Add your secrets in .env:
# NVIDIA_API_KEY=nvapi-xxxxxxxx
# NVIDIA_API_BASE=https://integrate.api.nvidia.com/v1/tts   # or your Riva/NIM endpoint
npm run dev

Visit http://localhost:3000

Test the integrations

• Text proxy: POST /api/nvidia (UI: /nvidia-test)
• TTS proxy: POST /api/nvidia-tts (UI: /nvidia-tts-test)

---

📦 Build & Start

npm run build
npm start

---

🛠 Development Scripts

• npm run lint – Lint the project with ESLint.
• npm run test – Run automated tests (if available).
• npm run build – Build the production bundle.
• npm start – Start the production server.

---

🧱 Tech Stack

• Next.js 14 + React 18 + Tailwind CSS
• GPT-OSS for context-aware phrase suggestions
• NVIDIA Speech AI / Riva for expressive voice synthesis
• dotenv for secure secret management

---

🛡️ Security

See SECURITY.md for responsible disclosure guidelines.

---

🤝 Contributing

We welcome contributions! Please review the CODE_OF_CONDUCT.md and CONTRIBUTING.md before submitting pull requests.

---

📽️ Demo Video

▶ 3-Minute Video Demo (coming soon)

---

🏆 Competition Categories

Submitted for:

• Best Overall – Innovative, high-impact use of GPT-OSS.
• For Humanity – Empowers marginalized voices.
• Most Useful Fine-Tune (future) – Fine-tuning on autism-specific AAC data.

---

🧭 Additional Docs

• GPT-OSS Agent Setup

---

🛡️ License

Licensed under the Apache License 2.0. See LICENSE.