Inspiration
We built Buddy because we wanted a gentle nudge to check in with ourselves, something warmer than a spreadsheet and lighter than a journal, a buddy that listens, spots patterns, and cheers the small wins without feeling clinical.
What it does
Buddy guides daily check-ins with mood sliders and customizable prompts, stores your data per-user in Firebase, and turns it into AI-powered insights for patterns, wins, and blind spots, with streaks and reminders to keep you engaged.
How we built it
The app runs on React + Vite + TypeScript and Tailwind css for styling, uses Firebase Auth and Firestore for secure, per-user data, and calls OpenRouter/OpenAI for AI summaries; local state is keyed by uid to keep prompts and settings isolated per account. Strict backend prompt engineering is used to eliminate hallucinations.
Challenges we ran into
We had some major issues with connecting our LLM as we thought it was the LLM issue, which took a lot of time debugging, and a chunk of the time, but it was
Accomplishments that we're proud of
We’re proud that we built a genuine tool, not a one-trick pony. With editable prompts/questions, Buddy can be whatever you need it to be a daily reflection companion, a mental health check-in, a fitness or habit coach, or simply a space to pause and just journal your day down! Instead of forcing a rigid workflow, we designed Buddy to adapt to your goals and evolve with you, so it grows more useful the more you make it your own.
What we learned
Working with LLMs and also learning how to get structured output from LLMs
What's next for Buddy
Getting more structured responses. Adding a calendar integration. Having a mobile app instead of a web app.
Built With
- firebase
- firestore
- openrouter
- react
- typescript
Log in or sign up for Devpost to join the conversation.