Argent

About the project

Inspiration Our financial lives are a mess. We have banking apps, investment apps, credit card statements, and budgeting spreadsheets—a scattered collection of data points that create more anxiety than clarity. These tools are just data museums; they show you a rearview mirror of what you've already spent, but offer zero wisdom on what to do next. The inspiration for Argent came from a simple, powerful question: What if you had a single source of truth for your entire financial world, and a dedicated team of specialists to help you understand it? We wanted to build a true Financial Headquarters—a platform that doesn't just show you data, but helps you build a strategy.

What I Learned The biggest lesson was that "AI" is not a monolith. Our initial idea of a single "chatbot" felt generic and quickly hit its limits. The true breakthrough was realizing that, like in the real world, you need a team of specialists. This led to the creation of our AI Trio: A Text Analyst for precision and data investigation. A Voice Advisor for speed and on-the-go convenience. A Video Strategist for "big picture" insights and scenario planning.

I learned that the user experience for each of these assistants had to be radically different to be effective. I also dove deep into Supabase's Row-Level Security to ensure that user data is not just functional, but fundamentally private and secure from the ground up.

How I Built It

Argent was built from the ground up with a modern, scalable architecture.

The Foundation: The project started with designing a robust PostgreSQL schema in Supabase. This included creating tables for profiles, accounts, and transactions, and then locking it all down with strict Row-Level Security (RLS) policies. Supabase Auth handles user authentication seamlessly.

The Frontend: The user interface was built using Next.js and React, developed inside the incredibly fast Bolt.new environment. Styling was done with Tailwind CSS to create the sleek, modern "Soft UI" aesthetic. The entire application is deployed globally via Netlify.

The X-Factor (The AI Trio): This is the core of the project.

The Text Analyst is powered by the OpenAI API (GPT-4), with system prompts engineered to provide detailed, formatted financial analysis.

The Voice Advisor uses the ElevenLabs API to provide instant, natural-sounding audio responses for quick queries. The Video Strategist is a conceptual framework designed to generate scripted, data-driven video briefings, combining narration with synchronized on-screen charts and visuals.(Thanks to TAVUS). 21st.dev - UI Deployment - Entri was flawless AI Integration - PicaOS

Challenges I Faced

Preventing AI Hallucinations: The biggest challenge was ensuring the AI assistants didn't provide dangerous or incorrect financial advice. This was solved by engineering highly specific system prompts that create "guardrails," forcing the AI to act as an analyst and strategist (describing data and running scenarios) rather than a prescriptive advisor (telling users what to buy or sell).

UI/UX for Three AIs: How do you integrate three different assistants without cluttering the interface? The solution was the floating, vertical "AI Hub," a persistent but unobtrusive element that provides instant access to the entire AI team from anywhere in the app. Thanks 21st.dev!

Database Security: Implementing RLS correctly was a challenge. It required careful planning to write policies that ensured a user could only ever access their own data, which is non-negotiable for a financial application.

Built With

• 21st.dev
• api
• auth
• deployment:
• elevenlabs
• netlify
• next.js
• platform:-bolt.new-frontend:-react
• row-level-security)-ai-apis:-openai-(gpt-4)
• tailwind-css-backend-&-database:-supabase-(postgresql
• tavus