Prodex AI

Prodex AI automates product creation from images, generating descriptions and tags, while enabling fast, scalable listings for ecommerce platforms

Comment 1

Inspiration

We were inspired by the growing demand for automation in e-commerce and retail, where businesses struggle to scale product listings and manage content efficiently. As the global e-commerce market surges, we saw an opportunity to simplify the process of product creation from images, leveraging AI to automatically generate product descriptions and enhance the product discovery experience for users.

What it does

ProdexAI automates the product creation process. Users can upload an image, and the system generates detailed, SEO-optimized product descriptions, tags.. It also supports image-based search, allowing users to find similar products using visual cues. All data is stored in a vector database, enabling retrieval-augmented generation (RAG) for intelligent product suggestions and queries.

How we built it

We built ProdexAI using a combination of cutting-edge tools and technologies to streamline the process of generating product descriptions from images. We used AWS Bedrock to manage infrastructure and ensure scalability. The core functionality relies on Llama 3.1 running through Ollama for generating product descriptions, supported by LlamaIndex for indexing and retrieval-augmented generation (RAG). All data, including product descriptions and image metadata, is stored in MongoDB Atlas, enabling fast retrieval and intelligent query handling. We also utilized a vector database to support image-based search, allowing users to find visually similar products. Our architecture is designed to handle real-time queries and generate descriptions quickly, ensuring a smooth and efficient user experience.

Challenges we ran into

Time constraints: With only 3 hours to build, managing and integrating all components (LLMs, vector database, RAG) quickly was a challenge. Optimizing search: Making the visual search smooth and ensuring it accurately retrieves relevant results using the vector database.

Accomplishments that we're proud of

Successfully implemented a fully functioning workflow where an image can be transformed into structured product data in real-time. Built a system that reduces the manual labor involved in product creation and marketing content, giving businesses a scalable solution.

What we learned

What's next for Prodex AI

Incorporate more sophisticated pricing and tagging mechanisms based on market trends and user preferences. Expand capabilities: Introduce multilingual support for global sellers and improve integration with popular e-commerce platforms (e.g., Amazon, Shopify). Real-time analytics: Add a dashboard for businesses to track how products perform and optimize listings further.

Share this project:

Updates