Skip to main contentDermUnbound

Flagship Project

The Docker Framework

A curated Docker-based framework for downloading, running, and managing open-source AI tools tailored for dermatology. One command to deploy an entire clinical AI suite on your own hardware.

Framework in Active Development — Target Release: Q3 2026

The architecture below describes our target deployment. Individual components are at various stages of development.

Target Architecture

Up and running in two commands

Illustrative output — Docker image not yet published.

terminal
# Pull the DermUnbound AI toolkit (planned)
$ docker pull dermunbound/derm-ai-toolkit

# Start all services
$ docker compose up

✓ Dermoscopy AI model loaded (ResNet-50, fine-tuned)
✓ Clinical documentation engine ready
✓ Pathology correlation service started
✓ Local inference engine initialized (GPU detected)
✓ Dashboard available at http://localhost:8080

The framework will download pre-configured Docker containers with all dependencies, models, and configuration files. No manual setup required. Planned support for Windows, macOS, and Linux.

Overview

What is the Docker Framework?

The DermUnbound Docker Framework is an open-source collection of containerized AI tools designed specifically for dermatology practice. Each tool runs inside a Docker container — an isolated, reproducible environment that works identically on any machine.

For the non-technical physician: think of it as a single application that bundles multiple AI tools together. You install Docker once, run one command, and get access to dermoscopy AI, surgical planning tools, clinical documentation generators, and pathology report processors — all running locally on your computer.

For the technically inclined: the framework uses Docker Compose to orchestrate multiple microservices. Each tool is independently versioned, tested, and deployable. The architecture supports GPU-accelerated inference via NVIDIA Container Toolkit and provides a unified REST API for integration with existing clinical systems.

Why Docker

Why containerized deployment?

One-command setup

No dependency hell. No version conflicts. One command downloads, configures, and starts everything.

Reproducible environments

The exact same setup on every machine. If it works on your laptop, it works on your clinic workstation.

Air-gapped operation

After initial download, no internet connection required. All processing happens locally.

Easy updates

Pull the latest version with a single command. Roll back instantly if needed. Zero downtime.

Planned Tools

What's inside the framework

Dermoscopy AI

Open-source skin lesion classification using publicly available ISIC-trained models. Local inference on dermoscopic images for melanoma, BCC, SCC, and benign lesion detection.

Surgical Planning

Defect analysis and reconstruction option generation for Mohs surgery. Algorithmic flap and graft recommendations based on defect geometry and anatomical location. Decision support only.

Clinical Documentation

Template-based clinical note generation. Structured reports for surgical procedures, consultations, pathology correlation, and follow-up documentation.

Pathology Correlation

Automated processing of pathology reports — text extraction, screenshot capture, patient-level organization, and clinical-histological correlation.

Local Inference Engine

GPU-accelerated model inference running entirely on your hardware. Supports ONNX, PyTorch, and TensorFlow models. No internet required after initial setup.

Privacy Architecture

Complete air-gap capability. All data processing happens on-premise. No telemetry, no cloud calls, no external dependencies during operation.

Requirements

System requirements

OSWindows 10/11, macOS 12+, Ubuntu 22.04+
RAM16 GB minimum, 32 GB recommended
Storage50 GB free (models + containers)
GPUNVIDIA GPU with 8 GB+ VRAM (optional, recommended)
DockerDocker Desktop 4.x or Docker Engine 24+

Privacy

How your data stays local

Every component of the Docker Framework is designed to operate without any external network connection. Once the containers are downloaded, the system functions entirely offline.

No telemetry. No usage tracking. No model phone-home. No cloud API calls. Patient images, clinical notes, and diagnostic outputs never leave your machine. The planned framework will include network audit tools to verify zero external communication.

Important Notice

The DermUnbound Docker Framework provides clinical decision support tools for research and educational purposes. It is not a medical device. It has not been cleared or approved by the FDA, CE, or any regulatory body. All clinical decisions remain the sole responsibility of the treating physician. These tools are not a substitute for professional clinical judgment.

Follow development

The Docker Framework is open source and free to use. Star the repo to follow progress.