The Poor People App.

The Foundation

The infrastructure that serves the most people in America is also the most underfunded and the slowest to update. 2.9 million federal civilian employees work inside systems that predate the smartphone. 19 million veterans navigate VA processes that average 220+ days for a disability decision. 3.5 million state and local health and social service workers are coordinating care on spreadsheets and in silos. 900,000 first responders operate within networks that have no shared coordination layer. And 53 million unpaid family caregivers are managing it all quietly — without pay, without proper tools, and without a system that was ever actually built for them.

There is a real chance that we can build our way into equivalent leverage — knowledge is now powered by AI. The infrastructure is being built to learn. Assets and decentralized coordination integrate rather than replace. It’s cleaner, and we minimize loss while we build the amplification of our network.

Two Things Everyone Has Heard Of

Think Uber’s research and coordination — and Tinder’s profile-centric matching service. Both are anywhere and anytime. We have learned to trust both to provide a service in a predictable manner. Uber allows for timely and predictable rides and ride shares. Tinder allows us to use images and descriptions to analyze whether we trust someone enough to start a conversation. PPA takes that same logic and applies it to something that matters even more: getting the right help to the right person at the right time, within a community that already knows each other.

PPA connects people who need help with people who can provide it — cutting bureaucratic overhead by a wide margin. Families with vetted caregivers. Seniors with companions. Anyone buried in forms or bureaucracy with navigators who know the system. The marketplace is broad by design — because the needs of a community don’t fit neatly into categories.

No waiting lists. No humiliation. Visual profiles and reputation-based matching where the best providers surface organically. Security keeps humans in the loop at every critical junction — because trust is the foundation, not a feature. We are going to use the knowledge we have as humans to build technology that helps us navigate our current digital evolution together.

The AI Agent Architecture

The architecture is intentionally model-agnostic — we’re building the harness, not betting on any single LLM. When the next frontier release ships, our agent structure adjusts to the relevant upgrades. Frontier and open-source models plug in where they perform best. The infrastructure doesn’t get rewritten — it gets better.

Future-proofing this network is an engineering requirement. We don’t yet know what models like Fable, Mythos, or other frontier and open-source models will be capable of in the coming months. What we do know is that a solely centralized platform might be a single point of failure. So we distribute: identity lives on Nostr — a protocol, not a company, which means no single entity can own, censor, or shut down the network. Community ownership through a token structure designed to keep value circulating within the network rather than extracted by intermediaries. Open-source intelligence pipelines that learn from public data, regulatory changes, and the collective knowledge of every Sentinel, caregiver, and neighbor in the system.

PPA is a platform for communities that want to organize, care for each other, and build resilience against the kind of institutional failures that are already in progress. The AI manages the coordination layer that most professionals must charge extra for because of administrative overhead. Human Sentinels handle safety and real judgment. AI is not ready to make decisions for us, but it can definitively help us coordinate and help us make better ones. The conceptual framework remains valid and the harness is designed to be ready for any major shift.

Claude · Hermes (NousResearch) · Frontier Models · Nostr · Bitcoin · Lightning · Open Source

Decentralized by Design

PPA must be able to fail while the community survives. Identities, reputation, and connections live on Nostr — a protocol, not a platform, which means no single company can own or shut down the network. Bitcoin backs the treasury. Lightning handles micro-payments. If the company dissolves, the network keeps running.

This isn’t just philosophy — it’s engineering for resilience. As AI systems grow more powerful and bad actors gain access to AGI-level tools, centralized platforms become single points of failure. A distributed network with no master database to breach, no central authority to compromise, and no single infrastructure dependency to attack is not a feature — it’s a survival requirement. We’re building for the world that’s coming, not the one that existed when most platforms were architected.

The Sentinel Program — Trust as Infrastructure

Every coordination network lives or dies on trust. PPA’s answer is the Sentinel Program: recruiting the people communities have always relied on — retired military, law enforcement, firefighters, and EMTs — to serve as verified trust anchors within local networks. Not for symbolism. Because they have spent careers keeping people safe, and that track record is the most valuable thing a platform like this can build on.

The relationship is genuinely reciprocal. Military families — active, veteran, and the families that military participants come from — are disproportionately represented among the communities PPA is designed to serve. They need elder care for parents, childcare during deployments, economic navigation, and community connection. They are both a primary user population and, through the Sentinel program, a primary source of the trust that makes the entire network function.

The same architecture that serves a military family in need of eldercare coordination can serve a VA outreach program, a nonprofit serving veterans’ families, or any institutional partner working across the care ecosystem. The infrastructure doesn’t need to be rebuilt for each context — it adapts. That’s the harness at work.

Full vision, agent layers, and the Sentinel Program live at poorpeople.app

OpenCare — The Live MVP

OpenCare is a care coordination platform built to solve a problem millions of caregivers face: coordinating care across people who don't use the same tools, don't live in the same place, and don't always communicate well. A caregiver creates a Care Circle — a shared workspace for everyone involved — then assigns tasks (medical, insurance, household, financial, legal, transportation), sets recurring schedules, and keeps everyone informed through automated email summaries and Telegram digests.

The platform features government ID verification through Persona, a REST API that AI agents use to manage care tasks via natural language, daily Telegram and email digests, and a progressive web app installable from any browser. Built on Next.js, TypeScript, Supabase, Clerk, and Docker — deployed on AWS with SSL and production-hardened security.

This is what the Poor People App architecture looks like in practice — shipped, working, free, helping people right now.

Whether you’re coordinating care for your family, running services inside a nonprofit, managing clinical intake for a health system, or building financial navigation tools for underserved communities — this infrastructure was designed to scale to your mission. The systems are evolving to serve families in every community across America, and beyond.

Build With Us

If you're a developer, operator, or organization that wants to work with this infrastructure — this is where it starts.