Skip to content

Installation

Prerequisites

  • Ollama — Local LLM runtime (ollama.com)
  • Python 3.12+ — For the backend (if running from source)
  • Node.js 22+ — For the frontend (if running from source)

The easiest way to run Confidential Translator — no coding knowledge required.

# Clone the repository
git clone https://github.com/jajupmochi/confidential-translator.git
cd confidential-translator

# Start with Docker Compose (includes Ollama)
docker compose up -d

Then open http://localhost:8000 in your browser.

GPU Acceleration

For NVIDIA GPU support, uncomment the deploy section in docker-compose.yml to enable GPU passthrough to Ollama.

Option 2: Standalone Binary

Download the latest release for your OS from the Releases page.

Platform Download
Linux confidential-translator-linux-amd64
Windows confidential-translator-windows-amd64.exe
macOS confidential-translator-macos-x64
  1. Download and run the executable
  2. The app will auto-start Ollama and open your browser to http://127.0.0.1:8000

Option 3: From Source

# Clone
git clone https://github.com/jajupmochi/confidential-translator.git
cd confidential-translator

# Build frontend
npm install -C frontend
npm run build -C frontend
cp -r frontend/dist/* backend/app/static/

# Install & run backend
cd backend
uv sync
uv run python -m app.main