## Inspiration

The soul of BIOSTRIDE was born from a convergence of two worlds and a shared dream. My girlfriend, Jenny, a Clinical Laboratory student, has always been deeply passionate about the field of prosthetics. She even earned a prestigious scholarship to specialize in this area, but due to harsh economic realities, she was unable to fund the remaining costs to attend.

Instead of letting that dream fade, we chose to iterate on it together. By merging her medical perspective with my background in Computer Engineering, we realized we could bridge a massive gap. Professional prosthetics are often a luxury, costing thousands of dollars. We want to empower individuals and animals to regain their mobility by providing them with the intelligence to design their own bionic solutions for the price of a single roll of 3D printer filament. This project is our way of standing together—medicine and technology—to support those who need it most.

## What it does

BIOSTRIDE is an AI-powered prosthetic design studio that democratizes access to mobility. The process is streamlined for the real world:

  • Input: A user uploads a single photo of their residual limb (human or animal) alongside a common reference object (like a coin or ruler).
  • Process: A team of AI agents powered by Gemini 1.5 Pro analyzes the anatomy and calculates scale, while Gemini 1.5 Flash reasons through material selection and physics.
  • Output: The platform generates a medical-grade STL file and a personalized assembly guide.

It acts as an essential tool for patients needing affordable starts and as a "Proof of Concept" platform for medical professionals to prototype clinical cases rapidly.

## How we built it

We engineered a sophisticated agentic workflow using LangGraph to manage the "reasoning loops" between our specialized agents:

  • Vision Analyst Agent: Uses multimodal vision to detect anatomical landmarks and calculate the scale ratio \(S\).
  • Engineering Agent: Translates measurements into 3D geometry using the numpy-stl library.
  • Safety Auditor: A reflective agent that verifies structural integrity and ergonomic safety before delivery.

To ensure medical-grade accuracy from a 2D photo, we calculate the scale ratio \(S\) using a reference object with a known diameter:

$$ S = \frac{\text{Actual size of reference (mm)}}{\text{Detected size in pixels (px)}} $$

The backend is powered by FastAPI, ensuring that the transition from a mobile upload to a downloadable 3D model happens in under 60 seconds.

## Challenges we ran into

The primary hurdle was depth perception and perspective. Translating a 2D image into a 3D coordinate system requires high precision to avoid ill-fitting sockets. We spent significant time on prompt engineering to ensure the Vision Agent could account for camera lens distortion and correctly identify reference objects in varying lighting conditions.

## Accomplishments that we're proud of

We successfully built a system that treats mobility as a universal right. We are particularly proud that our architecture supports both human and animal anatomy—helping a pet walk again is just as vital as helping a human regain independence. Achieving a professional-grade STL generation in less than a minute using only a smartphone photo is a milestone we are incredibly proud of.

## What we learned

Building BIOSTRIDE taught us that the true power of AI lies in "Agentic Workflows." By allowing different models to check and criticize each other’s work, we moved from a simple chatbot to a reliable engineering tool. We also learned that the most impactful technology is built at the intersection of empathy, medical knowledge, and computation.

## What's next for BIOSTRIDE

We plan to launch a Professional Dashboard for doctors to refine AI-generated meshes and a community-driven open-source library where users can share mechanical designs for personalized printed sockets, making bionic technology truly open for everyone.

Built With

Share this project:

Updates