An offline-first, conversational FinTech ledger designed for informal traders in emerging markets. Built over a fast-paced 12-hour hackathon track.

• 🎙️ Voice-to-Ledger: Speak local transactions in English or Swahili. Natural language fields are isolated directly into operational data structures using gemini-1.5-flash.
• 📷 Photo-to-Text OCR: Point-and-shoot physical purchase tracking. Processes blurry or handwritten market receipts instantly into structural expense variables via Gemini Vision intelligence.
• 📊 Visual Performance Analytics: Interactive bar charts tracking real-time local thresholds using high-contrast, outdoor-readable visibility guidelines via fl_chart.
• 💾 Offline Resilience Infrastructure: Built with continuous client functionality. Active network drop monitoring retains data safely on local disk structures using persistent Firebase indexing, auto-syncing upon reconnection.

• Frontend UI Framework: Flutter 3.24+ (Dart)
• State Management Matrix: Flutter Riverpod
• AI Engine: Google Gemini API via Google AI Studio
• Persistent Caching Database: Firebase Realtime Database (Spark Tier Configuration)

Create a .env file in the project's root folder:

GEMINI_API_KEY=YOUR_FREE_STUDIO_API_KEY_HERE