This is a submission for the Google AI Studio Multimodal Challenge
What I Built
I built a Micro-Business Digital Assistant that helps small business owners keep track of their daily sales and expenses with minimal effort.
The app provides:
Sales & Expense Tracking with manual entry and AI automation.
Persistent storage in the browser (IndexedDB) so data never disappears.
Summaries & Charts to visualize financial health
Multi-language support (English + Bengali) with local currency customization
The goal was to create a lightweight, offline-first tool that works on any modern browser without requiring sign-ups, servers, or external databases.
Demo
🔗 Live Applet on Cloud Run: https://micro-business-assistant-263910167686.us-west1.run.app/
🖼️ Screenshots:
How I Used Google AI Studio
I used Google AI Studio Build mode with Gemini 2.5 Flash to handle all multimodal inputs directly in the browser:
Image input: Photos of receipts or handwritten notes are parsed into structured JSON.
Audio input: Voice memos describing sales/expenses are transcribed and extracted into records.
Text summarization: Daily summaries are generated in English or Bengali with totals and a short narrative.
Multimodal Features
Image → Data (OCR): Upload a receipt/invoice/handwritten note → AI extracts type, item, quantity, amount, and date.
Audio → Data (Speech-to-Text + NLP): Record a quick note like “Sold 3 coffees for 150 taka today” → AI transcribes and saves it as a sale.
Summaries → Narrative: Daily totals computed locally, then AI generates a localized business summary in English or Bengali.
These multimodal features make the app accessible for business owners who may not have time to type every transaction manually. They can snap a photo or record a quick voice note.
Core Features
Sales & Expense Tracking: Log and categorize all transactions.
Persistent Storage: Records saved in IndexedDB for persistence.
Manual Data Entry: Add/edit records by hand.
AI Automation: OCR for receipts, speech-to-text for audio.
Interactive Dashboard: Summaries with charts and filters.
Data Table: Paginated, with live search.
Exports: CSV and Excel reports.
Customization: English/Bengali, custom currency.
Toast Notifications and Sample Data for smooth onboarding.
This project shows how Google AI Studio’s multimodal capabilities can be used to make business tools more natural and human-friendly — turning everyday receipts and voice notes into actionable financial data.
Top comments (2)
Feels like having a pocket accountant powered by AI..
Good.