Home Page
Features Section
Voice Feature Preview (viewers can try without login)
Image Import Preview (viewers can try without login)
Reviews Secction
Login Page
Signup Page
User Dashboard (after login)
Settings Page (of user logged in)
Daily logs (in user dashboard)
Command Center (works with voice and keyboard)
Handwritten Sheet to Inventory (loading status)
Handwritten Sheet to Inventory (completed analyzing)

Inspiration

During our research, we observed a simple but widespread reality among small retailers: many shopkeepers still manage their inventory using physical notebooks. For them, the notebook is the inventory system.

While this approach works for daily operations, it becomes difficult when records need to be tracked over longer periods, corrected, searched, or analyzed. Updating stock manually is time-consuming, and maintaining accuracy becomes increasingly challenging as inventory grows.

We wanted to create a solution that felt natural for retailers rather than forcing them to learn complex software. That idea became Vocalize — a voice-first inventory assistant that allows shopkeepers to manage stock simply by speaking.

What it does

Vocalize is an AI-powered inventory management platform designed specifically for small retailers.

Instead of manually entering stock data, retailers can simply speak commands such as:

"Add 10 bags of rice"
"Sold 3 notebooks"
"Set water bottles to 50"

The system intelligently determines whether the user is having a conversation or issuing an inventory command.

If the input is conversational, Vocalize responds naturally like an assistant.

If the input is an inventory command, Vocalize extracts the required information, updates inventory records, and generates an audit trail of the action.

To improve usability, the most recently modified inventory rows are automatically highlighted, allowing retailers to instantly identify what changed without searching through their inventory.

Vocalize also supports AI-powered handwritten inventory digitization. Users can upload handwritten stock sheets, invoices, or inventory records, and the system converts them into structured inventory data. Before importing, every extracted field remains editable, allowing users to review and correct any OCR mistakes.

Additional features include:

Voice-first inventory management
Conversational AI assistant
Handwritten stock sheet digitization
Inventory audit logs
Multi-language dialect support
Product search and filtering
Editable inventory management
AI-assisted product matching

How we built it

We built Vocalize as a full-stack web application.

Frontend:

React
TypeScript
Tailwind CSS
Web Speech API

Backend:

Node.js
Express.js

Database:

MongoDB Atlas
Mongoose

Artificial Intelligence:

Google Gemini

The voice workflow begins with speech recognition in the browser. The spoken transcript is then sent to Gemini, where a two-stage AI pipeline determines whether the user is making conversation or performing an inventory operation.

For inventory commands, Gemini extracts structured actions and inventory data. For conversations, Gemini generates contextual assistant responses.

For document imports, Gemini is used to analyze uploaded handwritten inventory sheets and transform them into structured product records that can be reviewed and imported into the inventory database.

Challenges we ran into

One of the biggest challenges was ensuring that natural speech could be reliably converted into inventory operations.

Retailers do not speak in a fixed format. Different users describe inventory updates differently, making command interpretation difficult.

Another challenge was preventing duplicate inventory entries caused by spelling variations, abbreviations, and product naming inconsistencies.

Building a seamless handwritten-sheet-to-inventory workflow was also challenging. OCR extraction is never perfect, so we designed an editable review interface that allows users to validate and correct extracted information before importing it.

Finally, we focused heavily on user experience. Rather than forcing users to search for modifications, we automatically highlight recently updated inventory records to provide immediate visual feedback.

Accomplishments that we're proud of

Building a complete voice-first inventory management experience
Creating an AI system that distinguishes conversations from commands
Developing handwritten inventory digitization with editable verification
Designing a retailer-friendly workflow that requires minimal technical knowledge
Implementing visual change tracking through automatic row highlighting
Supporting multiple spoken language dialects for broader accessibility

What we learned

This project taught us valuable lessons about AI-powered user experiences, prompt engineering, voice interfaces, OCR workflows, inventory management systems, and designing technology for users who may not be comfortable with traditional software.

We also learned that successful AI products are not just about model capabilities—they are about creating workflows that users can trust and easily understand.

What's next for Vocalize

Our vision is to evolve Vocalize into a complete AI operating system for small retailers.

Future plans include:

Regional language expansion
Supplier and purchase management
Inventory analytics and forecasting
WhatsApp-based inventory updates
Sales trend insights
Multi-store inventory synchronization
Mobile applications for Android and iOS
Offline-first capabilities for low-connectivity regions

We believe technology should adapt to people, not the other way around. Vocalize is our step toward making inventory management as simple as having a conversation.