Inspiration
Smriti was inspired by a deeply personal experience — my grandmother has Alzheimer’s. Watching her slowly lose not just memories, but emotional connections and a sense of identity made us realize that existing solutions focus mostly on reminders and tasks, not on preserving how moments felt. We wanted to build something that could help people living with Alzheimer’s stay emotionally connected to their memories and loved ones, while also supporting caregivers who often feel overwhelmed and unsure.
What it does
Smriti is an AI-powered memory and emotion companion for people living with Alzheimer’s and dementia. It allows users to capture daily moments through voice or text and automatically stores them with emotional context. When users feel confused or forgetful, Smriti helps them recall memories through natural, empathetic conversations. For caregivers, Smriti provides emotional insights and trends over time, enabling more informed and personalized care.
How we built it
We built Smriti using a combination of speech-to-text, natural language processing, emotion analysis, and conversational AI. User inputs are converted into context-rich memory representations that link people, events, time, and emotions. These memories can later be retrieved through a conversational interface. We also built a caregiver dashboard that visualizes emotional patterns and highlights potential concerns. Throughout development, we focused on accessibility, privacy, and a voice-first experience suitable for elderly users.
Challenges we ran into
One of the biggest challenges was dealing with the subjective nature of emotions. Emotions cannot be accurately inferred from a single interaction, so we designed Smriti to focus on emotional patterns over time rather than one-off predictions. Another challenge was balancing technical ambition with the limited time of a hackathon, which required us to carefully prioritize features that would create real-world impact without overcomplicating the system.
Accomplishments that we're proud of
We’re proud of building a working prototype that goes beyond traditional reminder systems and focuses on emotional memory. Successfully integrating emotion-aware recall with a caregiver insight system within a short hackathon timeframe was a major achievement. Most importantly, we’re proud that Smriti is grounded in empathy and real human needs, not just technology.
What we learned
Through this project, we learned that building AI for healthcare requires more than technical accuracy — it demands empathy, ethical thinking, and user trust. We gained a deeper understanding of how emotional data should be handled carefully and how involving caregivers in the loop improves both safety and impact. We also learned the importance of clear communication when presenting complex ideas.
What's next for Smriti AI
Next, we plan to validate Smriti through real-world pilot testing with caregivers and early-stage Alzheimer’s patients. We aim to improve multilingual and voice support, refine emotion tracking with human feedback, and explore partnerships with healthcare organizations and Alzheimer’s support groups. Our long-term vision is to expand Smriti into a broader emotional memory and cognitive wellness platform that supports patients, families, and caregivers over time.
Built With
- mern
- mistral
- pinecone
- python
- tensorflow
Log in or sign up for Devpost to join the conversation.