Quickstart Guide

Get the Vision system up and running in under 10 minutes.

Prerequisites

1. Hardware:
   • NVIDIA GPU (Pascal architecture or newer).
   • Minimum 4GB VRAM.
2. Software:
   • Docker Desktop or Docker Engine.
   • NVIDIA Container Toolkit.
   • Git.

Installation

1. Clone the Repository

git clone git@github.com:FirstBreath/firstbreath-vision.git
cd firstbreath-vision

2. Configure Environment

Copy the example configuration if present, or create .env from scratch:

# If .env.example exists in the repo root:
cp .env.example .env

# Otherwise create .env and set at least: DB_MAIN_*, DB_CAMERA_*, REDIS_*
# See the [Environment Variables](/docs/vision/configuration/environment) reference.

Edit .env to set your MySQL and Redis credentials (hosts must match your backend stack when using Docker network control-hub-network).

3. Build & Launch

Use Docker Compose to build the stack. The first build effectively compiles the TensorRT engine, which can take 5-10 minutes.

docker-compose up -d --build

Verification

Check the logs to ensure the system is healthy:

# check camera manager
docker logs -f control-hub-camera-manager

# check inference
docker logs -f control-hub-batch-inference

You should see messages like Connected to Redis and Inference engine cold-booted.