Venditor AI
Venditor AI is an AI-powered assistant that helps small vendors turn WhatsApp chat screenshots into organized business orders automatically.
Inspiration
Many small vendors run their businesses entirely through WhatsApp. Orders are often scattered across chats, making it difficult to track customers, manage deliveries, and stay organized.
I wanted to build a simple solution that helps vendors organize their business without forcing them to learn complicated software or leave the platforms they already use daily.
That idea led to Venditor AI.
What it does
Venditor AI allows vendors to upload WhatsApp screenshots containing customer orders.
Using MeDo’s multimodal large language model capabilities, the app:
- Extracts order information from screenshots
- Converts unstructured chats into organized order data
- Saves orders into a dashboard
- Generates AI-powered customer replies
- Provides smart business suggestions based on order activity
The platform is designed to feel simple, fast, and accessible for everyday vendors.
How we built it
Venditor AI was built using MeDo and its multimodal AI capabilities.
We used:
- MeDo official large language model
- OCR and image understanding workflows
- AI-powered structured JSON extraction
- Local storage for demo persistence
- Modern responsive UI generation with smooth animations
The application flow works like this:
- Vendor uploads a WhatsApp screenshot
- OCR extracts visible text
- MeDo AI processes the extracted content
- The AI returns structured order data
- Orders are saved and displayed in the dashboard
We also implemented:
- AI-generated WhatsApp-style replies
- Smart business insights and recommendations
- Interactive dashboard experience
- Smooth animations and micro-interactions
Challenges we ran into
One of the biggest challenges was ensuring the AI returned structured and consistent order information from messy chat screenshots.
We solved this by carefully structuring prompts and forcing the AI to return clean JSON responses.
Another challenge was balancing functionality with simplicity. The goal was to create something powerful while still feeling lightweight and easy for vendors to use.
Accomplishments that we're proud of
- Building a complete AI-powered workflow using MeDo
- Successfully converting screenshots into structured orders
- Creating a clean and intuitive vendor experience
- Adding AI business suggestions for vendors
- Delivering a polished prototype within hackathon constraints
What we learned
This project showed how multimodal AI can solve practical business problems for informal vendors and small businesses.
We also learned how important product simplicity and user experience are when building AI-powered tools.
What's next for Venditor AI
Future plans include:
- Multi-user support
- Inventory tracking
- Voice-based order logging
- Cloud database storage
- Multi-language support
Built With
- bome
- framer
- gemini
- lucide
- medo.ai
- pnpm
- postgresql
- react
- row-level
- security
- shadcn
- superbase
- tailwind
- typescript
- vite
Log in or sign up for Devpost to join the conversation.