Skip to main content

VirtueGuard-Code Local Setup Tutorial

This guide walks you through deploying VirtueGuard-Code locally using Docker with GPU support.

Prerequisites​

Before you begin, ensure you have the following installed on your system:

  • Docker installed and running
  • CUDA installed with GPU support enabled

Setup​

1. Pull the Docker Image​

Pull the VirtueGuard-Code backend image from the container registry:

docker pull us-west1-docker.pkg.dev/customer-docker-virtueai/customers-poc/virtueguard_code_backend:latest

2. Tag the Image​

Tag the image for easier reference:

docker tag us-west1-docker.pkg.dev/customer-docker-virtueai/customers-poc/virtueguard_code_backend:latest virtueguard_code_backend:latest

3. Run the Container​

Start the container with GPU support:

docker run -d \
  --name virtueguard_code_backend \
  --gpus all \
  -p 4409:4409 \
  -e ANTHROPIC_API_KEY="your_apikey" \
  -e OPENAI_API_KEY="your_apikey" \
  -e JWT_SECRET_KEY="NmCxDRTwSbaABkkb2b4y4UW46S8lZf9R" \
  virtueguard_code_backend

Note: ANTHROPIC_API_KEY and OPENAI_API_KEY are optional environment variables. They are only required if you want to use Claude or GPT models for comparison purposes.

4. Wait for Model Deployment​

Wait approximately 20 minutes for the model to download and deploy. The backend needs time to initialize the vulnerability detection model.

5. Get the Default API Key​

Once the container is running, retrieve your API key from the logs:

docker logs -f virtueguard_code_backend

You should see output similar to this:

Docker logs showing API key

Copy the API key — you will need it for the next step.

Usage​

1. Install the Extension​

Install the VirtueGuard-Code extension from the marketplace of VS Code, Cursor, or Windsurf.

VirtueGuard-Code extension installation

2. Configure the Settings​

Open the extension settings and make the following changes:

VirtueGuard-Code extension settings

  • API Base URL: Update to http://localhost:4409 (this points to your local Docker container)
  • API Key: Paste the API key you copied from the Docker logs (e.g., sk-vai-d63fe7774f55eaff34e99901479731d1e50d436e1b7a16dad45f201dcf5951cb)

3. Start Scanning​

You can now start scanning code with the local model! Follow the usage instructions in the VirtueGuard-Code Dashboard Tutorial for details on:

  • Autoscan Mode: Automatic vulnerability detection on file save
  • Manual Scan Mode: Targeted analysis of selected code sections

Troubleshooting​

API connection issues​

Verify the container is running and the port is accessible:

docker ps
curl http://localhost:4409/health

Model not loading​

Check the container logs for errors:

docker logs virtueguard_code_backend