DermUnbound is a research group developing open-source, privacy-first clinical AI tools for dermatology. Every tool is designed to run locally on the physician's own hardware, eliminating cloud dependency and ensuring patient data never leaves the clinic.

DermUnbound is led by Dr. Yehonatan Kaplan, a board-certified dermatologist and Fellowship-trained Mohs micrographic surgeon based in Israel. The project is a self-funded research initiative combining clinical dermatology expertise with software development.

A clinician-coder is a practicing physician who uses AI-assisted development tools to build their own clinical software. Rather than relying on technology companies to interpret clinical needs, the clinician-coder translates domain expertise directly into working tools using modern AI-assisted coding workflows.

Yes. All DermUnbound tools are released as open-source software. Source code is publicly available on GitHub, allowing any physician or developer to inspect, audit, modify, and redistribute the tools. Transparency is a core principle of the project.

The Docker Framework is DermUnbound's deployment architecture that packages clinical AI tools into self-contained Docker containers. Each container includes all dependencies, models, and interfaces needed to run a tool with a single command, requiring no manual configuration or technical expertise.

Most DermUnbound tools run on standard consumer hardware. A modern computer with 16 GB of RAM and 20 GB of free disk space is sufficient for most tools. Specific tools that run large language models locally may benefit from additional RAM or a dedicated GPU, but this is documented per tool.

A GPU is not required for most DermUnbound tools. Tools that leverage large language models can run on CPU-only hardware, though a CUDA-compatible NVIDIA GPU will significantly improve inference speed. Each tool's documentation specifies whether GPU acceleration is supported.

Updates are delivered as new Docker container versions. To update, you pull the latest container image and restart the tool. There are no forced updates or automatic patches. You control when and whether to update, and you can always roll back to a previous version.

No. DermUnbound tools process all data locally on the physician's machine. No patient data, images, or clinical information is transmitted to external servers. The tools are designed to function completely offline with zero cloud dependency.

Air-gapped means the software can operate on a computer that has no internet connection at all. DermUnbound tools are air-gap capable, meaning once the Docker container is downloaded, the tool runs entirely offline with no network access required for any functionality.

Local AI processes all data on the physician's own hardware. Unlike cloud AI services that require uploading patient data to external servers, local AI ensures that sensitive clinical information never leaves the physician's control. This eliminates risks of data breaches, unauthorized access, and cross-border data transfers.

Yes. Because DermUnbound tools are open source, you can inspect the complete source code to verify there are no network calls or telemetry. Additionally, running the Docker containers in air-gapped mode with no network access provides a hardware-level guarantee that no data can be transmitted.

MohsPedia is an AI-powered knowledge base for Mohs micrographic surgery. It provides instant access to procedural guidance, flap and graft references, and surgical decision support. MohsPedia runs locally and is designed to assist Mohs surgeons during clinical practice.

DermTools is a comprehensive digital toolkit for dermatologists, available at dermai.co.il. It includes structured clinical documentation templates, an interactive differential diagnosis wizard, dermoscopic algorithms (2-Step, 7-Point, Chaos & Clues), clinical scoring calculators (PASI, SCORAD, DLQI), a topical steroid ladder with Israeli trade names, drug dosing and interaction tools, patient education handouts in Hebrew, ICD-10 codes, laboratory monitoring guides, interactive DDx matrices, and Israeli health basket drug status tracking.

OptiMohs is a decision-support tool for Mohs surgery case selection. It helps surgeons evaluate whether a given skin cancer case is appropriate for Mohs micrographic surgery based on tumor characteristics, anatomic location, and established clinical guidelines.

PathCorrelate is a clinical-pathological correlation tool that helps dermatologists match clinical presentations with histopathological findings. It supports the diagnostic workflow by providing structured correlation between what is seen clinically and what is observed under the microscope.