Memories That Care

MemoryCare is more than a chatbot, it's a compassionate AI companion for patients. It remembers family information, manages medications, celebrates goal completion, and adapts over time.

Comment

Inspiration

Our inspiration came from witnessing the daily struggles faced by people with dementia or Alzheimer's. With over 55 million people worldwide living with memory-related conditions, we saw a critical gap: existing AI chatbots forget everything between sessions, making them useless for those who need consistency and continuity. We thought to ourselves, what if AI could actually remember? What if it could recall a patient's family members, medication schedules, and personal preferences across weeks and months? That's when we realized we could build something truly transformative, an AI companion that doesn't just chat, but genuinely knows its users.

What it Does

MemoryCare is a compassionate AI health companion powered by MemMachine's persistent memory system. It serves two main user types:

For Patients:

  • Conversational AI Companion: Engages in warm, natural conversations that reference past interactions and personal details
  • Medication Tracking: Reminds patients about medications with daily dose tracking and adherence monitoring
  • Goal Management: Tracks therapeutic goals assigned by doctors and automatically detects completion through natural conversation
  • Dual Memory System: Maintains both episodic memories (events, conversations) and profile memories (permanent facts about family, hobbies, preferences)

For Healthcare Providers:

  • Patient Management Dashboard: Monitor multiple patients from a single interface
  • Medication Oversight: Prescribe medications, set schedules, and view adherence data
  • Goal Assignment: Set therapeutic goals and track patient progress over time
  • Memory Inspection: Review patient interactions and stored information

The key innovation is that everything persists. Restart the app, log out and back in weeks later and MemoryCare still remembers every detail about each patient.

How we built it

We started by getting a working chatbot that can store information about users and filter it by preferences and other categories. Next we worked on improving the UI while simultaneously improving our filter and also including better prompts for the AI. Then we added the medication and goals dashboards, along with making the AI keep track of both through natural conversation.

Challenges we ran into

We had a few challenges with user info cross contamination and getting the UI to sync with our back-end code for things such as the dashboards, but thanks to persistence and caffeine we were able to overcome our challenges.

Built With

Share this project:

Updates