Project summary

MedScale is an existing, independently maintained, free Apache-2.0 local-first desktop health AI project with FHIR-native workflows and Windows/macOS engineering-beta builds. Its public code and project information are at https://github.com/AbdulazizShehri/MedScale and https://medscale.si/. The project is not clinically certified, and the engineering tests described here use fictional/synthetic data only. I am seeking voluntary grant funding to make the software more trustworthy and maintainable, not to introduce paywalls.

Goals and approach

Over a proposed six-month funding period, I aim to:

1. Expand the reproducible, synthetic-only FHIR evidence-attribution, abstention, privacy and reliability test suite; publish methods and results that other developers can run independently.

2. Obtain an independent external security review and publish a responsible summary and remediation evidence without disclosing unpatched weaknesses.

3. Improve Windows/macOS build and regression testing, documentation, and offline-first usability, with a particular focus on devices available to independent maintainers.

4. Publish regular progress updates, transparent limitations, and an openly accessible reproducibility guide. These are intended deliverables, not completed or promised clinical outcomes.

What success would look like

A public Apache-2.0 testing harness and synthetic fixtures; repeatable scripts and reports; an independently reviewed security improvement plan and evidence of fixes; tested Windows/macOS engineering builds and a documented maintenance process. The results will be measured by reproducible pass/fail cases, verified code changes and publicly reviewable artifacts, not by unverified patient usage or medical effectiveness.

How funding would be used

The full proposed funding goal of $25,000 USD is a planning budget: $10,000 for maintainer engineering time, $7,500 for an independent security assessment and verification, $5,000 for development/test hardware, and $2,500 for documentation, accessibility and reproducibility testing. Quotes and actual vendors would be selected transparently after funding; they are not signed contracts. If only the $5,000 minimum is raised, I would reduce scope to a documented synthetic evaluation and a bounded independent security review, and not promise full delivery of all milestones. Funding will be used only for public-benefit open-source work, not for patient data or paid clinical services. No expense will be double-charged against any other grant.

Team and track record

This is currently an independent maintainer-led project. I have an MPH from the University of Melbourne, an AWS certification, and I maintain public MedScale code, build/test documentation and engineering-beta artifacts. The repository is the best evidence of prior work. I am not claiming institutional sponsorship, a clinical deployment, patient adoption, or a named external reviewer who has not agreed to participate.

Risks, limits and transparency

Risks include hardware availability, time needed for security review, dependency vulnerabilities, compatibility issues, and a limited maintainer capacity. If funding is lower than hoped, I will narrow milestones, report the actual results and remaining gaps, and follow the eventual grant agreement for any unspent funds. The public project remains free and independently controlled. No health records from real patients will be used in the grant work.

Existing funding

No cash grant, device shipment, or GPU allocation has been confirmed for MedScale at the time of posting. Other grant inquiries and applications are pending; this proposal does not count them as money received. If more than one funder supports the work, separate cost accounting will prevent double funding.

Writing disclosure

This factual proposal was organized with AI-assisted drafting; the maintainer is responsible for verifying the project description, budget assumptions and deliverables.