-
Main Dashboard
-
Displaying the streaks and dynamic results
-
AI generated results and also resilient fallback mode
Inspiration
People journal to process emotions, but most entries remain unstructured and don’t lead to action. We wanted to build a privacy-first tool that turns reflection into practical next steps.
What it does
MindLedger is an AI-powered journaling companion. For each entry, it returns:
- mood score and emotional label
- trigger detection
- emotional insight
- personalized affirmation
- one 5-minute micro-habit
It also generates a weekly report with:
- average mood and trend
- emotional theme
- pattern detection and top triggers
- 7-day action plan with completion tracking
- trigger rescue plans with concrete interventions
How we built it
We built MindLedger with React + Vite and a lightweight UI focused on clarity and speed. Data is stored in localStorage (no backend, no accounts). AI uses Gemini as primary provider with fallback logic for reliability.
Challenges we ran into
- API quota limits and model availability issues
- getting consistent structured JSON from model responses
- keeping the app reliable when cloud AI is unavailable
- balancing rich features with a simple user experience
Accomplishments that we're proud of
- end-to-end polished product with clean UX
- actionable wellness flow (insight -> habit -> progress)
- resilient fallback architecture
- privacy-first design with local data storage
- live deployment ready for judging
What we learned
- reliability matters as much as AI quality
- users need actionable guidance, not only analysis
- focused scope beats feature bloat in hackathons
- privacy-first architecture can still deliver strong value
What's next for MindLedger
- mobile-first version
- report export options
- deeper personalization and trend intelligence
- stronger wellness safety guardrails
Built With
- anthropic-api-(optional-fallback)
- css
- gemini-api
- html
- javascript
- localstorage
- react
- vite
Log in or sign up for Devpost to join the conversation.