Somewhere in El Paso right now, a family is doing the math. Their mother is getting older. The house is getting harder to manage. A nursing home feels wrong — she raised four kids in that kitchen, she knows every crack in the tile. But leaving her alone feels impossible too.

This is the gap we are building for. Not the gap between young and old, but the gap between the care families want to give and the tools that exist to help them give it.

The problem with existing solutions

The default answer has been the wearable panic button — a device the elder wears that lets them call for help if they fall. It is a reasonable idea. It also does not work.

Elders refuse wearables. They find them stigmatizing, uncomfortable, or simply forget to put them on. The device that is supposed to save them sits on the nightstand. Meanwhile, generic smart home platforms — Google Home, Amazon Alexa, Apple HomeKit — are built for convenience, not care. They automate lights and thermostats. They do not understand that a person who normally walks to the kitchen by 7:30am has not moved in two hours.

The family caregiver is left managing pure anxiety — calling constantly, driving over to check, missing work, missing sleep, missing the signal when it finally matters.

Old home values, supplemented by a hyper-sensored agentic AI that does a better job at minimizing signal to noise — like grandma's chicken soup.

— Treevine Life, on what CLASP is supposed to feel like

Introducing CLASP

CLASP — the Caregiver Logistics, Assistance, and Support Platform — is a passive AI monitoring system for in-home, non-medical palliative care. It is built on three principles that distinguish it from everything that came before.

Principle one — passive by design

CLASP asks nothing of the elder. A network of sensors, cameras, and smart speakers already present in the home — or added as modular, interchangeable hardware — monitors movement, activity, appliance use, and voice. The elder does not wear anything. Does not press anything. Does not change a single habit. The system works around them.

Principle two — acts, not just alerts

When a stove is left on for thirty minutes with no one in the kitchen, CLASP does not send a notification and wait. A relay physically cuts the power. No internet connection required. No cloud call. The action executes locally, on a small hub device inside the home, regardless of what the network is doing. A voice reminder follows in Spanish: "Mija, la estufa lleva un rato encendida." The family caregiver gets a push alert. The whole thing happens in seconds.

Principle three — learns this person, not just people

Every elder is different. When she normally wakes up. How often she visits the kitchen. Which nights she sleeps through and which nights she does not. CLASP builds a behavioral baseline specific to this person in this home — and it gets smarter every day. The first anomaly it catches might be a changed routine. The fiftieth might be an early indicator of a health change that has not yet shown up clinically.

Why privacy is built into the architecture

The three-zone model is not a compliance checkbox. It is a design constraint that shaped every technical decision we made. The world's best AI — Google Gemini — never sees a patient's name, face, address, or medical history. It receives anonymized questions about behavioral patterns and returns structured observations. The patient's private data lives encrypted on a small hub device inside their own home. It does not leave.

This matters for two reasons. First, it is the right thing to do. Second, it is what allows CLASP to get smarter over time without creating a surveillance system. Every alert a caregiver dismisses or escalates becomes a labeled data point that improves the model — for that home, and eventually for the broader pattern of what elder care in the Borderplex actually looks like. No one else is building that dataset.

Built for the Borderplex, from the Borderplex

Treevine Life is an El Paso company. Our first prototype is being built for an El Paso family. Spanish is not a feature we added — it is the default. The voice prompts, the alert messages, the conversational interface — all of it is authored in Spanish first, in the regional dialect of the US-Mexico border, not machine-translated from English.

The Borderplex is one of the largest bi-national metropolitan regions on the continent. It has a multi-generational family caregiving culture, a fast-growing elderly population, and dramatically lower adoption of technology-based care solutions. That is not a problem. It is a market.

What comes next

The first prototype is being built now. The hardware ecosystem runs on the Matter protocol — the open smart home standard backed by Google, Apple, Amazon, and Samsung — which means any certified sensor, camera, lock, or speaker can be a CLASP node. We are building with Google Home and Nest hardware and integrating Google's full AI stack: Gemini for real-time inference, Vertex AI for model training, and Agent Studio for the conversational care interface that will eventually live in a Google Home speaker with Gemini.

The software subscription launches first. A family can onboard to CLASP without purchasing a single piece of hardware. The sensors and actuators come later, as the relationship deepens. That is how we stay a software company — and how CLASP gets to sit in homes long enough to become genuinely useful.

We are applying to the ELEVATE HealthTech Challenge at the MCA Innovation Center. The Borderplex is our launch market, and we intend to be the platform that defines what AI-assisted elder care looks like in bilingual, multi-generational households — starting here, scaling from here.