We believe healthcare should move at the speed of science, not bureaucracy. Regulatory challenges don't just delay life-saving technologies; they also inhibit market entry for new medical devices and prevent promising innovations from ever leaving the research lab. To help address this, we are building a set of collaborative, AI-powered tools to cut the cost of regulatory compliance, speed up market access, and empower medical technologists to invest in innovation, not paperwork. Our first step is OpenReg, an open-source tool that will streamline the FDA premarket approval process for startups, research teams, and other medtech innovators working on medical devices.

The Problem: An Unnecessary Bottleneck

The FDA’s premarket approval process for devices is a critical bottleneck for medical innovation. Even its new digital submission format, eSTAR, creates unnecessary barriers. The system is built on a proprietary, PDF-based format that resists collaborative workflows and lacks crucial compliance features. This is like a locked, opaque filing cabinet that hides the rules. Teams are forced to manually copy data and email partial files, increasing the risk of costly errors and rejections.

This status quo is extremely unfavorable to small companies. While larger firms can and do employ whole teams of regulatory specialists, early-stage companies are primarily reliant on expensive consultants to demystify their path to market, with document preparation alone costing $20,000 to $50,000. OpenReg seeks to solve this inequality by turning publicly available rules into open-source, machine-readable software, allowing innovators to invest in innovation, not paperwork.

The Solution

We’re building tooling to make the FDA submission requirements modular, machine-readable, and interoperable—so startups, universities, consultants, and software vendors can all build on the same compliance backbone. This design ensures OpenReg is not another proprietary silo, but a shared public standard.

OpenReg will be an open-source, AI-powered compliance platform for medtech innovators, with our development phased to maximize impact at each funding milestone:

• The Foundation (Core MVP): Our initial deliverable will be the core engine—an eSTAR PDF parsing tool and an interactive web app. This foundational layer unlocks collaboration by transforming the opaque FDA logic into a transparent, machine-readable format (JSON model), immediately reducing manual data entry errors.

• The Regulatory Co-pilot: The next stage integrates an AI Compliance Assistant. This regulatory co-pilot provides real-time feedback and intelligent suggestions as teams build their submission, preventing costly rejections and saving hundreds of person-hours in the review process.

• The Generative Platform: At its full scope, OpenReg becomes a generative tool capable of automatically drafting entire submission sections from your source materials. This final stage also integrates a compliant Quality Management System (QMS), which dramatically accelerates time-to-market by automating a significant portion of regulatory work.

The Impact

We envision a future where regulatory compliance is a seamless, automated extension of the product development lifecycle. We will measure OpenReg’s impact by tracking reductions in time and cost for FDA submissions, improvements in submission quality, and breadth of adoption across the medtech ecosystem. Key metrics include time from draft to submission-ready, person-hours and consulting costs saved, first-pass acceptance rates, and reductions in Additional Information requests. We will also measure the diversity and number of organizations deploying OpenReg, integrations with existing compliance tools, and open-source community contributions. Over time, we will assess broader systemic effects such as faster time-to-market, increased participation from first-time or academic-origin teams, and improved transparency and traceability in regulatory workflows—indicators that OpenReg is lowering barriers and accelerating safe, effective medical innovation.

The Team

Our team’s expertise spans regulatory and go-to-market strategy, statistical modeling and data science, LLM application development, and data infrastructure. We are uniquely positioned to deliver a modular, open-source compliance platform that integrates seamlessly into medical software workflows.

Our collaboration began at Harpocrates, an AI strategy and software services firm serving investors, venture-backed startups, and academic research teams.

• Bryan Davis, our CEO, brings over a decade of experience of building machine learning powered software and large-scale data infrastructure, as well as previous entrepreneurial experience as cofounder and CTO of a venture-backed legaltech startup, which specialized in document processing for legal workflows.

• Logan Stokols, our CTO, is a former research professor of mathematics at Duke University who has a background in high-dimensional data analysis and partial differential equations (PDEs). His work since has focused on ML and AI systems for medical technology and quantitative finance.

In our work, we have architected software for multiple regulatory and compliance frameworks, including ISO 27001, SOC2, HIPAA and ISO 13485. We have encountered many innovators who began to consider regulatory and compliance constraints only after significant work had been completed, resulting in them needing to backtrack on key decisions, change significant implementation details, and integrate compliant version-control procedures in already completed systems. Small companies in particular face a dual challenge in the regulatory process: not only do they lack the infrastructure to build and document with regulation in mind, they also operate under intense capital constraints, where unexpected delays can be potentially fatal for their businesses.

We first discovered the particular problem of FDA compliance for medical devices in our work with a UNC School of Medicine research team who sought to spin out their research into a commercial venture. As we assisted this team of established clinicians and professors in transforming their work into market-ready medical software, we saw how intertwined product development and regulatory and go-to-market strategy were. Working with the team as they faced the information vacuum of the regulatory affairs industry, dominated by expensive consultants and antiquated software, was a catalyst for our conviction that new tools were needed.

Our community

Our project is supported by advisors with deep expertise in medical device regulation and digital health innovation. One advisor has 14 years of experience at the intersection of federal health agencies and the medical device industry, including leadership roles shaping innovation policy, regulatory science programs, and cross-sector public-private partnerships. Another served for over a decade in senior positions at the U.S. medical device regulator, including founding leadership of its digital health initiative, where they helped set national policy for AI/ML-enabled medical technologies and other emerging health software.

We also already have deep ties to our target users: early-stage founders and research teams seeking market entry for their AI-enabled medtech innovations. Our customers include early stage founders working on a new wearable for monitoring blood glucose levels, a team working on a new form of AI-augmented psychometrics, and several academic research groups at Duke and UNC actively seeking to transform their research into products.

We believe we can make a significant impact on human well-being by building open-source LLM-powered tools that make the FDA application process much more efficient and transparent, thereby democratizing and speeding access to medical breakthroughs. We have seen firsthand how regulatory bottlenecks delay life-saving technologies, and we know the potential impact of using language models to significantly improve document review and preparation. With our combined experience and access to outside experts, we are confident in our ability to build OpenReg into the backbone of a system for faster, safer FDA applications, arming small teams with new tools to navigate this landscape efficiently.

Why now?

The confluence of two major trends makes the present moment uniquely opportune for OpenReg. Firstly, the exponential growth in Software as a Medical Device (SaMD) and AI-driven diagnostics is creating unprecedented regulatory volume and complexity. The traditional, paper-based submission infrastructure is buckling under the weight of continuous software updates and the novel risk profiles of AI. Simultaneously, the FDA's recent push toward digital standards, like the eSTAR program, signals an urgent attempt to modernize, but its proprietary, PDF-based implementation fundamentally resists the collaborative, data-driven workflows essential for modern software development. This growing gap—a surge in SaMD complexity met with an antiquated, closed-source compliance tool—creates an immediate and critical need for an open-source, machine-readable infrastructure like OpenReg to handle the unique demands of digital health.

Secondly, the rapid maturation of large language models (LLMs) provides the necessary technological leverage to tackle this escalating regulatory challenge for the first time. Regulatory compliance is fundamentally a document and knowledge management challenge, perfectly suited for the generative and analytical capabilities of modern AI. OpenReg capitalizes on this timing to move beyond simple document parsing and build an intelligent "Regulatory Co-pilot" that can understand context, provide real-time compliance feedback for complex SaMD requirements, and even draft submission sections. This capability was simply not feasible until recently. By integrating this AI power into an open framework today, OpenReg can fundamentally democratize access to sophisticated regulatory expertise, making it a sustainable public good for the rapidly expanding digital health ecosystem rather than an expensive, proprietary service accessible only to the largest firms.

Our fundraise

Without this work, the application of AI to the FDA submission process will almost certainly be through proprietary, closed-source tools. The status quo will continue, where large firms can afford expensive solutions, while small companies and research teams are left with few options to navigate the path to market for their new medical technologies. There will be little progress toward creating the public standards and open-source tooling needed to raise the bar for all players.

Our vision for OpenReg is to create a suite of regulatory tools that are open, accessible, and intelligent. We’re at the very beginning of our journey, and our ambition over the next 6-18 month period is to build as much of this as possible.

• $100,000 (The Foundation): This funding level delivers the Core Minimum Viable Product (MVP), which includes the eSTAR PDF parsing engine and an interactive web application for building submissions. This development unlocks collaboration and significantly reduces manual data entry errors for early-stage teams.

• $200,000 (The Regulatory Co-pilot): This level expands the platform by adding the AI Compliance Assistant, which provides real-time feedback and suggestions. This feature prevents costly rejections and streamlines the review process, ultimately saving hundreds of person-hours.

• $500,000 (The Generative Platform): At this final stage, OpenReg evolves into a generative tool. It will be capable of automatically drafting entire submission sections and will integrate a compliant Quality Management System (QMS). This comprehensive automation accelerates time-to-market for new medical devices by taking over a significant portion of the regulatory work.