Actively building · March 2026

Custom prosthetics for
children mothers families survivors communities a better future

Three cofounders building a scan-to-print prosthetics pipeline that delivers custom sockets at 80–90% cost reduction. Follow our open journey to deploy the first prosthetic to Gaza.

📱 Intake App→ 🤖 AI Validation→ 📐 3D Scan→ 🖥️ CAD→ 🖨️ 3D Print→ 🦿 Patient

Scroll to explore

The Problem

Prosthetics are broken
where they're needed most

The global prosthetics market is structurally broken: devices are expensive, slow to deliver, impossible to customize at scale, and virtually inaccessible in conflict zones.

$5K–50K

Cost per prosthetic device
(plus repeated refits)

Weeks

To months of wait time
for a single device

70%

Of conflict amputations
are lower limb

In Gaza, thousands of amputees need prosthetics now — not in months. The traditional model can't scale. We're building a digital pipeline so any field worker with a phone and a 3D printer can deliver a custom-fit socket.

What We've Built

Real code, real decisions,
real progress

Not mockups. Not wireframes. Working software that runs on a phone and a laptop.

📱

LimbScanner — Patient Intake App

An iPhone app for field workers to capture and validate limb photos before a prosthetist visits. Smart object segmentation and skin detection enable quick remote intake — reducing wasted trips and enabling triage from anywhere.

12-angle guided capture with progress ring and directional chevrons
Works on any iPhone — no LiDAR or special hardware needed
On-device quality checks in <20ms (sharpness, brightness, contrast)
Real-time AI analysis over Wi-Fi — segmentation mask, confidence score, overlay toggle
Per-frame saving — no data loss if the app crashes mid-capture
Limb type selection: foot, hand, forearm (left/right)

🧠

AI Limb Detection Engine

A FastAPI backend powered by RMBG-2.0 (0.2B parameter BiRefNet model) that segments limbs from any angle — including residual limbs and close-up shots that defeated YOLO and MediaPipe.

Foreground segmentation + HSV skin-tone validation
No skeleton context needed — works on amputated anatomy
~1 second inference on CPU (no GPU required)
REST API: iOS app sends JPEG, receives mask + overlay + metrics
EXIF orientation handling for iPhone photo rotation
CLI tools for batch validation of entire scan folders

🤖

Agent-Powered Operations

Six specialized AI agents manage our entire project — from grants to meetings to competitor research. The repo is the company. Auto-syncs to Google Drive for non-technical collaborators.

Meeting Agent — prep, notes, action items, follow-ups
Grant Agent — discovery, eligibility checks, application drafting
Research Agent — competitor deep-dives, technical research
Pitch Agent — deck content generation, audience tailoring
12 formal engineering decisions logged with full context
Document verification system — every file tracks last review date

Custom prosthetics for children mothers families survivors communities a better future