Video Transcription: Myrror is an application that allows users to get a genuine reflection of who they are and connect using this with their peers. Myrror runs in the background of your computer following your online presence and filtering the data through a three stage RAG pyramid (basic chunking, session chunking, overall summary chunking) which is then analyzed by a local Ollama model to gauge your big 5 personality, give things you could possibly be interested in, and tell you your spirit animal. With your generated personality we find the people in our database that you are most likely to connect with. Our website is hosted on vercel and the personalities are stored in a Supabase backend.
Inspiration: We wanted to turn everyday digital breadcrumbs into a privacy-first self-reflection engine that makes personal data feel transparent, useful, and human.
What it does: Myrror transforms consented browsing, search, and feed exports into an on-device behavioral intelligence dashboard with interest maps, signal scoring, rhythm analysis, and a personal data portrait.
How we built it: We built Myrror using a python based back-end, with a local-first web app using HTML, CSS, and JavaScript, with a client-side ETL pipeline, consent-gated ingestion, signal normalization, and interactive data visualization.
Challenges we ran into: The biggest challenge was balancing powerful psychographic-style insights with privacy-by-design guardrails, messy export formats, and fully browser-based processing.
Accomplishments that we're proud of: We’re proud that Myrror delivers a polished, AI-adjacent personal analytics experience without scraping, cloud tracking, invasive permissions, or third-party data pipelines.
What we learned: We learned that ethical data products can still feel intelligent and high-impact when consent, transparency, and local-first architecture are treated as core features.
What's next for Myrror: Next, Myrror can add richer export connectors, smarter semantic clustering, encrypted local storage, shareable insight reports, and optional on-device AI summarization.
Built With
- css
- html
- javascript
- python
- supabase
- vercel
Log in or sign up for Devpost to join the conversation.