Inspiration
It is hard to interpret the data I receive from my Credit Card .. like, I want to ask questions, and to get a summary of my month expenses clearly explained, some transactions are also weird (maybe fraud), I want them to be automatically to be detected.
What it does
A 3-part intelligence suite: Synthesizer: Auto-generates narrative monthly spending reports. Watchdog: Pushes real-time alerts for anomalies (Big errors, subscription surplus, geographic fraud). Chat Agent: Lets users ask natural-language questions directly to their ledger.
How we built it
Frontend: React + Tailwind CSS (deployed on Vercel). Backend: Python + Flask + SQLite. AI: Gemini API using precise Tool Use to query the database dynamically.
Challenges we ran into
Context window limits. We couldn't dump the whole database into the prompt, so we implemented LLM Tool Use to let the AI query specific SQL data only when needed.
Accomplishments that we're proud of
Building and deploying a full-stack app in one sprint. We're especially proud that the agentic part works and that what we wanted to have is working ! It will be an interesting product for Brim.
What we learned
A lot to discuss, a lot of AI agentic and deployment.
What's next for Financial Intelligence Dashboard
Maybe to scale the solution and make it work for multiple users.
Log in or sign up for Devpost to join the conversation.