• 🤖 AI Flashcard Generation — Powered by NVIDIA AI for text + Pollinations for save-time mnemonic images
• 🧠 Spaced Repetition System (SRS) — Hybrid SM-2 + Leitner algorithm for optimal memory retention
• 🏝️ Vocabulary Islands — Learn words in contextual topics (cooking, politics, tech, etc.)
• 📊 Smart Analytics — Track your learning progress, streaks, and mastered words
• 🔐 OAuth Authentication — Secure login via W9 Labs unified auth system
• 🎨 Modern UI — 2026 trending color palette with 8-bit voxel arcade elements

• Backend: Rust 1.94 + Axum 0.7 + tokio
• Database: PostgreSQL 16 (shared with W9 Labs ecosystem)
• AI: NVIDIA API for text generation + Pollinations for image generation
• Authentication: OAuth 2.0 via w9-db
• Deployment: Docker + Docker Compose + Caddy reverse proxy
• Frontend: Server-rendered HTML + vanilla JavaScript

# Clone the repository
git clone https://github.com/w9labs/vocai.git
cd vocai

# Copy environment file
cp .env.example .env

# Start PostgreSQL (or use existing w9-postgres)
# Create database: CREATE DATABASE w9_vocabai;

# Run the server
cd server
cargo run

docker compose up -d

vocai/
├── server/
│   ├── Cargo.toml          # Rust dependencies
│   └── src/
│       ├── main.rs         # Application entry point
│       ├── db.rs           # Database connection & migrations
│       ├── models.rs       # Data models
│       ├── handlers/       # Route handlers
│       │   ├── mod.rs
│       │   ├── auth.rs     # OAuth authentication
│       │   ├── flashcards.rs # Flashcard CRUD & generation
│       │   ├── islands.rs  # Vocabulary islands
│       │   └── review.rs   # SRS review sessions
│       ├── nvidia.rs       # NVIDIA AI integration
│       ├── srs.rs          # Spaced repetition algorithms
│       ├── auth.rs         # OAuth callback & login
│       └── session.rs      # Session management
├── public/
│   └── assets/
│       ├── css/vocai.css   # 2026 color palette + voxel theme
│       └── js/vocai.js     # Frontend interactions
├── Dockerfile              # Multi-stage Docker build
├── docker-compose.yml      # Docker Compose config
└── .env.example            # Environment template

Vocai uses a hybrid SRS combining:

1. SM-2 Algorithm (SuperMemo) — Calculates optimal review intervals based on recall quality
2. Leitner System — Moves mastered words to less frequent review boxes

Vocai deploys to the W9 Labs VPS alongside other services:

# SSH to server
ssh -p 22001 root@ffm.w9.nu

# Navigate to deployment directory
cd /opt/w9-labs

# Pull latest and redeploy
docker compose pull vocai && docker compose up -d vocai

Add to /etc/caddy/Caddyfile:

vocai.top {
    reverse_proxy vocai:3010
}

For static frontend:

wrangler pages deploy public --project-name=vocai

• users — User accounts (synced with w9-db OAuth)
• vocabulary_islands — Topic-based word collections
• flashcards — Individual vocabulary cards, image prompts, and Pollinations image URLs
• srs_reviews — Spaced repetition progress
• study_sessions — Review history
• user_stats — Aggregated learning statistics

Vocai uses the W9 Labs unified OAuth 2.0 system:

1. User clicks "Login with W9" on vocai.top
2. Redirect to db.w9.nu/oauth/authorize
3. User authenticates via w9-db
4. Callback to vocai.top/auth/callback with auth code
5. Exchange code for token via /oauth/token
6. Session created, user redirected to dashboard

1. The text model generates flashcards and an image_prompt for each card.
2. Pollinations renders the image only when the user saves a card.
3. Vocai stores the returned image URL plus prompt/model metadata in PostgreSQL.
4. If image generation fails, the flashcard still saves so study flow is never blocked.

• Primary: Electric Indigo (#6366F1)
• Secondary: Neo Mint (#A7F3D0)
• Accent: Digital Lavender (#E9D5FF)
• Background: Deep Space (#0F172A)
• Success: Cyber Green (#22D3EE)
• Warning: Solar Orange (#FB923C)
• Error: Crimson Glitch (#F43F5E)

• Headings: Space Grotesk (geometric, modern)
• Body: Inter (clean, readable)
• Code/Accent: JetBrains Mono (technical)

MIT License — see LICENSE for details.

Contributions welcome! Please read our Contributing Guide first.

Part of the W9 Network ecosystem. Learn more at w9.nu