ElevIQ — GenAI Personal Finance Assistant
A cross-platform (web + Android/iOS) AI finance assistant on a shared Firebase backend: photograph a receipt and Google Gemini extracts vendor, amount, date and line-items into categorised expenses, chat with an AI over your spending and uploaded statements, and manage goals, budgets, bills, net worth and 15 financial calculators — with no bank integration.
1 / 1Overview
Overview
ElevIQ is a GenAI personal-finance assistant delivered as both a Next.js web dashboard and a Flutter mobile app (Android/iOS) on one shared Firebase backend. It removes manual data entry by using Google Gemini to read receipts and analyse financial documents — with no bank-account linking.
Key Features
- AI receipt scanning — photo → Gemini Vision → categorised expense, using 7 category-specific extraction prompts with confidence scoring
- Streaming AI finance chat over your data and uploaded statements/bills (images, PDF, CSV)
- Full money toolkit — dashboard, spending analytics, goals, net worth, bills, reminders, spending limits and bill-splitting
- 15 working financial calculators — EMI, SIP, FIRE, old-vs-new tax regime and more, with verified formulas
- Guided onboarding for 8 user types, plus PDF/CSV export
- Custom auth — sequential human-readable IDs, email + Google account linking, and httpOnly session cookies
Architecture
Two native clients (React/TypeScript web, Dart/Flutter mobile) over a single Firebase project. Firestore stores expenses in a shared top-level collection keyed by userId for real-time sync across devices; Cloud Functions handle budget-alert triggers and scheduled summaries; Gemini is called server-side from Next.js API routes on web and on-device from the Flutter scanner.
Tech Stack
Web: Next.js 16, React 19, TypeScript, Tailwind 4, Zustand, Zod, Recharts, jsPDF · Mobile: Flutter, Dart 3, Riverpod, go_router, fl_chart, Dio · Backend: Firebase Auth, Firestore, Storage, Cloud Functions (Node 20) · AI: Google Gemini (2.5-flash, 3-flash-preview, 1.5-flash)
How it came together
Manual expense tracking is tedious and most finance apps require risky bank-account linking, so people abandon budgeting and lose visibility into where their money goes.
A cross-platform (web + Android/iOS) AI finance assistant on a shared Firebase backend: photograph a receipt and Google Gemini extracts vendor, amount, date and line-items into categorised expenses, chat with an AI over your spending and uploaded statements, and manage goals, budgets, bills, net worth and 15 financial calculators — with no bank integration.
Shipped a large real-time ecosystem — a ~23-screen web dashboard, a 27-page Flutter app, and a Firebase backend (Auth, Firestore, Storage, 7 Cloud Functions) — with core expense data syncing live across web and mobile.
Project timeline
Planning & Design
Feb 2026Feature spec, Firestore schema, and product/branding for the finance ecosystem.
Web MVP
Feb 2026Next.js dashboard with local state, laying out expenses, goals and analytics.
Database-driven Migration
Feb–Mar 2026Replaced local state with Firestore real-time hooks; added session cookies and the custom user-document layer.
AI Integration
Mar 2026Gemini receipt scanning, streaming chat, insights, and voice-to-expense API routes.
Flutter App
Mar 2026Feature-first Flutter client (Riverpod + go_router) reusing the Firebase backend and web API.
Backend & Deploy
Mar 2026Cloud Functions for budget alerts and cron summaries; deployed web to Vercel, Flutter to Android/iOS/Web.