Redefine

AI-powered word learning, reimagined.

Redefine is an intelligent vocabulary learning app that uses LLMs to generate rich explanations and Anki flashcards for any word, phrase, or concept.

✨ Features

• Personalized Learning — Everyone learns differently. Edit the prompts to generate definitions and flashcards the way you understand best—change the language, style, complexity, or format to match how your brain works
• AI-Powered Explanations — Get clear, contextual explanations with pronunciation, related concepts, and usage examples
• Smart Flashcard Generation — Auto-generates Anki-style cloze cards embedded in real-world sentences
• Multi-Provider LLM Support — Works with OpenAI, Anthropic Claude, and Google Gemini
• Anki Integration — Export flashcards to Anki with a single tap
• Mobile-First Design — Built for learning on the go. Instead of searching in your browser, search here and get flashcards you can import directly into Anki
• Lightweight — Uses only ~15MB of memory
• Self-Hosted — Your data stays with you

🚀 Getting Started

Option 1: Docker Compose (Recommended)

Create a docker-compose.yml file:

services:
  redefine:
    image: ghcr.io/kirarpit/redefine:latest
    container_name: redefine
    ports:
      - "5612:5000"
    environment:
      - SALT_KEY=${SALT_KEY}
    volumes:
      - redefine_data:/data
    read_only: true
    cap_drop:
      - ALL
    security_opt:
      - no-new-privileges:true
    restart: unless-stopped

volumes:
  redefine_data:

Create a .env file with a salt key:

# Generate a salt key and save to .env
echo "SALT_KEY=$(openssl rand -base64 32)" > .env

Start the app:

docker-compose up -d

Open http://localhost:5612 and add your LLM in Settings.

Important: Keep your .env file safe! If you lose the SALT_KEY, you'll lose access to your encrypted API keys.

Option 2: Docker Run

# Generate a salt key and save to .env
echo "SALT_KEY=$(openssl rand -base64 32)" > .env

# Run the container
docker run -d \
  --name redefine \
  -p 5612:5000 \
  --env-file .env \
  -v redefine_data:/data \
  --read-only \
  --cap-drop=ALL \
  --security-opt=no-new-privileges:true \
  --restart unless-stopped \
  ghcr.io/kirarpit/redefine:latest

Open http://localhost:5612 and add your LLM in Settings.

Important: Keep your .env file safe! If you lose the SALT_KEY, you'll lose access to your encrypted API keys.

🔧 Configuration

Adding an LLM Model

1. Navigate to Settings
2. Click Add Your First Model
3. Enter your model details:
   • Model ID: gemini/gemini-2.0-flash or gemini/gemini-2.5-flash-lite (recommended)
   • API Key: Your API key

Recommended: Gemini (Free)

We recommend gemini/gemini-2.0-flash or gemini/gemini-2.5-flash-lite — both are fast, high-quality, and have generous free tiers.

👉 Get a free API key from Google AI Studio

Other Supported Models

Provider	Format	Example
Google	google/<model>	gemini/gemini-2.0-flash
OpenAI	openai/<model>	openai/gpt-4o
Anthropic	anthropic/<model>	anthropic/claude-3-5-sonnet-20241022

Anki Integration

On Mobile (Recommended):

The easiest option is to tap "Download for Anki" and import the file directly into AnkiMobile or AnkiDroid.

For automatic syncing, you can run AnkiConnect on your phone—Redefine will detect it and enable "Send to Anki" for one-tap imports. (Note: AnkiConnect integration has been tested on mobile only.)

🛠️ Local Development

If you want to contribute or customize:

# Start the backend
cd server/go
./run.sh

# In a new terminal, start the frontend
cd web
npm install
npm run start

Open http://localhost:3000

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📄 License

This project is licensed under the MIT License.

---

Created with ❤️ for smarter learning