Caregiver Catena

Caregiver Catena is an AI-powered companion designed to support the millions of unpaid family caregivers who look after loved ones with Alzheimer’s at home. Caregivers are often isolated, overwhelmed, and untrained—yet they shoulder the majority of dementia care worldwide. The app provides an emotionally intelligent chat assistant that gives practical, empathetic guidance for daily challenges like agitation, confusion, resistance to care, routines, sleep issues, and communication.

It maintains a personalized care memory (diagnosis, living situation, preferences, routines), includes a private journal for tracking behaviors and emotions, and offers a simple routine/pill planner to reduce decision fatigue. Built on the 0G compute network for private, verifiable AI inference, CaretakerCatena supports caregivers without compromising their dignity or data.

The long-term vision is to evolve into a trusted global platform where anonymized in-home caregiving patterns become a valuable resource for science, healthcare systems, and families—helping improve early detection, care quality, and caregiver wellbeing at scale.

We built Caregiver Catena as a simple but powerful tool that brings real support to Alzheimer’s caregivers. The project has two main parts: a backend (the “brain” that talks to the AI model) and a frontend (the interface the caregiver sees).

The backend is a small server that connects to the 0G compute network. 0G is important because it lets us run AI models in a way that protects privacy and keeps costs low. Instead of sending private caregiving conversations to a big tech cloud, our server sends them to a 0G provider, which runs the AI model inside a secure environment. The backend handles setting up a wallet, checking the balance, choosing an available provider, and sending the caregiver’s message to the AI. One tricky part was learning how 0G authenticates each request using signed “one-time” headers, which we had to generate manually.

On the frontend, everything is done with vanilla HTML, CSS, and JavaScript so it runs anywhere without installation. We designed a clean one-column layout to keep things simple, especially for stressed caregivers. The UI includes a chat with typing animations, a “memory” section where we store what the caregiver tells us about their loved one, a local journal that saves entries on the user’s device, and a pill/routine planner to make daily care easier. No external database is used, which keeps the app private and lightweight.

We did a few hacky things to make the demo smooth: • We built a small “prompt builder” to combine the caregiver’s message, the conversation history, and the care profile so the AI always responds in context. • We added fallback logic so if the 0G provider fails, the app still returns a helpful message instead of crashing. • We made all the sections collapsible so the interface stays calm and uncluttered during presentations.

Overall, the project combines a sensitive user experience with a privacy-protecting AI backend, and it works entirely without storing personal data on a server. It’s simple on the surface, but there are a lot of thoughtful small details behind it.