dFDA (Decentralized Framework for Drug Assessment) is open-source software for ranking treatments by real-world outcomes and publishing an Outcome Label for each one. It is designed so patient records stay with patients and clinics and only aggregate results are shared; that sharing isn't built yet (see What works today below). It is an independent open-source project, not affiliated with, endorsed by, or acting on behalf of the U.S. Food and Drug Administration.

flowchart LR
  SAFE["Digital Twin Safe<br/>(patient's own data)"] -- "OAuth: share with my doctor" --> NODE["Clinic Node<br/>(runs at the clinic)"]
  NODE -- "Summary File<br/>(aggregates only)" --> AGG["Global Aggregator"]
  AGG --> SITE["dfda.earth<br/>Treatment Rankings + Outcome Labels"]

What works today

See Apps and features for the current application, planned packages, and features to extract from related repositories. The implementation sequence and data rules are in docs/MIGRATION.md, the authoritative implementation roadmap. The broader vision below is background, not a second backlog or a claim that every feature exists.

Billions of people are suffering needlessly because the current system of clinical research, diagnosis, and treatment sucks because:

• ⏳ Counterproductive Regulatory Barriers to clinical research block life-saving treatments by 7-12 years
• 🚫 97% of patients are excluded from clinical trials
• 💰 Drug development costs of $2.6B are passed on to patients
• ⏱️ Terminal patients wait 4+ years for breakthrough therapy approvals
• 📊 The system ignores real-world evidence about effective treatments

The Cure Acceleration Act creates:

• ✅ Universal Trial Access - Every person's right to try safe treatments
• 🤖 dFDA (Decentralized Framework for Drug Assessment) - Free, open infrastructure for real-world evidence collection
• 🏆 50/50 Health Savings Sharing Rewards - Multi-billion dollar incentives for developing actual cures instead of lifetime drug subscriptions
• 📈 Real-Time Analysis of the positive and negative effects of every food, supplement, drug, and treatment on every measurable aspect of human health and happiness
• 🌐 Global Access - Decentralized trials anyone can participate in from home

👉 Read the Full Cure Acceleration Act

The current system of clinical research, diagnosis, and treatment is failing the billions of people are suffering from chronic diseases.

👉 Problems we're trying to fix...

By harnessing global collective intelligence and oceans of real-world data, we hope to emulate Wikipedia's speed of knowledge generation.

This is a very high-level overview of the architecture. The three primary primitive components of the framework are:

1. Data Silo API Gateway Nodes that facilitate data export from data silos
2. Digital Twin Safes that import, store, and analyze your data to identify how various factors affect your health
3. Clinipedia that contains the aggregate of all available data on the effects of every food, drug, supplement, and medical intervention on human health.

Gateway API Nodes should make it easy for data silos, such as hospitals and digital health apps, to let people export and save their data locally in their Digital Twin Safes.

👉 Learn More About Gateway APIs

Digital Twin Safes are applications that can run on your phone or computer. They import, store, and analyze your data to identify how various factors affect your health. They can also be used to share anonymous analytical results with the Clinipedia FDAi Wiki in a secure and privacy-preserving manner.

Each Digital Twin Safe combines encrypted local storage with a personal AI agent that applies causal inference algorithms to estimate how various factors affect your health.

A local application for self-sovereign import and storage of personal data.

👉Learn More or Contribute to Digital Twin Safe

Personal AI agents that live in your Digital Twin Safe and use causal inference to estimate how various factors affect your health.

👉Learn More

The Clinipedia wiki should be a global knowledge repository containing the aggregate of all available data on the effects of every food, drug, supplement, and medical intervention on human health.

👉 Learn More or Contribute to the Clinipedia

A key component of Clinipedia is Outcome Labels that list the degree to which the product is likely to improve or worsen specific health outcomes or symptoms.

👉 Learn More About Outcome Labels

• Data Collection
• Data Import
• Data Analysis
  • 🏷️Outcome Labels
  • 🔮Predictor Search Engine
  • 🥕 Root Cause Analysis Reports
  • 📜Observational Mega-Studies
• Real-Time Decision Support Notifications
• No Code Health App Builder
• Personal AI Agent
• Browser Extension

Collects and aggregate data on symptoms, diet, sleep, exercise, weather, medication, and anything else from dozens of life-tracking apps and devices. Analyzes data to reveal hidden factors exacerbating or improving symptoms of chronic illness.

Web and mobile push notifications with action buttons.

By using the Browser Extension, you can track your mood, symptoms, or any outcome you want to optimize in a fraction of a second using a unique popup interface.

The Analytics Engine performs temporal precedence accounting, longitudinal data aggregation, erroneous data filtering, unit conversions, ingredient tagging, and variable grouping to quantify correlations between symptoms, treatments, and other factors.

It then pairs every combination of variables and identifies likely causal relationships using correlation mining algorithms in conjunction with a pharmacokinetic model. The algorithms first identify the onset delay and duration of action for each hypothetical factor. It then identifies the optimal daily values for each factor.

👉 More info about data analysis

More info about real time decision support

👉 More info about the predictor search engine...

👉 More info about observational studies...

The Crowdsourcing Cures site has its own repository. The retired fda-gov-v2 repository is a feature source for apps/web, not a second product line. Port useful behavior in reviewed slices; do not merge its entire divergent history into this repository.

The planned shared packages are listed in docs/MIGRATION.md.

• Frontend: React, Next.js, TypeScript, Tailwind
• Canonical node database: PostgreSQL through Supabase, with Row Level Security and SQL migrations
• Legacy data: a Prisma schema of the old MySQL database in packages/legacy-import, used only for the migration
• Authentication: Supabase Auth, plus the web app's own OAuth server for third-party apps
• Background jobs: graphile-worker
• Tooling: pnpm workspaces, Turborepo, Vitest, Playwright, GitHub Actions

• Node.js 22+
• pnpm
• Docker (for the local Supabase stack)
• Supabase CLI

1. Clone the repository:

git clone https://github.com/mikepsinn/dfda.git dfda
cd dfda

2. Install dependencies:

pnpm install

3. Set up environment variables:

cp apps/web/.env.example apps/web/.env
# Edit apps/web/.env with your Supabase credentials

4. Start the web app:

pnpm --filter web dev:env

This starts local Supabase (it needs Docker), then the Next.js server, the background worker (dev:worker) and the reminder cron (dev:cron). dev:next starts only the Next.js server, and dev starts it with secrets from Doppler instead of .env.

5. The first time, load the database schema and seed data from a second terminal:

pnpm --filter web db:local:reset

The web app uses Supabase for its database and authentication. From the repo root:

pnpm --filter web sb:local:start   # start only local Supabase
pnpm --filter web db:local:reset   # apply migrations and seeds
pnpm --filter web db:local:types   # regenerate TypeScript types

The database schema is managed through migrations in the apps/web/supabase/migrations directory. Each migration represents a specific change to the database structure.

The development environment includes:

• A local Supabase stack (PostgreSQL, Auth, Storage) started by sb:local:start
• A graphile-worker process for reminders and background jobs (dev:worker)
• A cron process that queues the periodic reminder jobs (dev:cron)

The repo uses pnpm workspaces and Turborepo, with the following structure:

dfda/
├── apps/           # Deployable applications (web)
├── packages/       # Shared libraries and tooling
├── schema/         # Earlier, unapplied database design
├── supabase/       # Earlier combined-migration tooling
├── docs/           # Documentation
└── .github/        # GitHub workflows

• pnpm --filter web dev:env - Start the web app with local Supabase, the worker and the cron
• pnpm --filter web build - Build it
• pnpm --filter web test - Run its tests
• pnpm --filter web lint - Lint it

1. Place the feature in apps/web or a shared package according to the migration plan
2. Create a new branch: feature/your-feature-name
3. Implement the feature following the architectural guidelines
4. Add tests and documentation
5. Submit a pull request

The reference web app deploys to Vercel from apps/web, using Supabase Cloud for its database and authentication. See the app environment setup for configuration.

Packaging the same app so a clinic can install and operate its own isolated Clinic Node is planned work. Follow the migration plan; there is no separate infrastructure deployment roadmap here.

Before nodes share data, implement consent records, authenticated node submissions, and the aggregate privacy controls. Public pages remain in apps/web; network administration belongs to the planned aggregator.

Use the development workflow above and the migration plan to choose and scope a contribution.

• Tested analysis methods and trial-results ingestion
• Patient tracking, OAuth/data imports, and shared export parsers
• Consent, privacy-preserving Summary Files, and node authentication
• Clinic Node packaging and aggregator administration
• Documentation, internationalization, and accessibility

Follow the implementation sequence: foundation, ratings and trial results, the dfda.earth cutover, the clinic network, and consent-based legacy migration. The feature inventory distinguishes existing code from extraction candidates and new work.

• Discover trials matched to your health profile
• Securely store and control your health data
• Track your participation and outcomes
• Receive personalized insights and recommendations

• Create and manage decentralized trials
• Access diverse patient populations
• Reduce administrative overhead
• Analyze real-time trial data

• Refer patients to appropriate trials
• Monitor patient participation
• Access comparative effectiveness data
• Integrate with existing EHR systems

• Build on the dFDA API platform
• Create specialized tools for clinical research
• Integrate with Gateway Nodes
• Preserve consent, access controls, and evidence provenance