AI E-commerce Assistant Development
Budget: $1,500 – $3,000 USD
I plan to launch an AI-driven assistant that simplifies marketing for sellers on Amazon, eBay, Shopify, Etsy, and Walmart. The first release combines a lightweight web app with a Facebook Messenger bot. Core features I need built right now are:
• Auto-generated product descriptions in Bangla, English, and additional languages.
• Auto-generated ad copy in the same language set.
• One-click AI photo editing for background removal and basic retouching.
A streamlined interface where users choose a marketplace, paste a few product details, and instantly receive polished copy and images is essential. Content must respect each platform’s style guidelines and export cleanly for direct upload.
While customer-query automation (WhatsApp/Messenger) and sales-insight modules will come later, the current architecture should leave room for those add-ons—modular code, clear APIs, and well-documented endpoints.
Acceptance criteria
1. A working prototype that handles at least 20 test products per marketplace without errors.
2. Language toggle producing grammatically correct Bangla and English; “Other languages” can route through an external translation API for now.
3. Image editor returning web-optimised PNG/JPG under 500 KB in <10 seconds on a mid-tier server.
4. Deployment guide and source code in a private repo, plus a short video walkthrough.
Technologies are flexible—Python, Node.js, or similar—with the option to plug in OpenAI or comparable LLMs for text, and an established vision API for images. Clear, maintainable code and concise documentation matter more than stack choice.
• Auto-generated product descriptions in Bangla, English, and additional languages.
• Auto-generated ad copy in the same language set.
• One-click AI photo editing for background removal and basic retouching.
A streamlined interface where users choose a marketplace, paste a few product details, and instantly receive polished copy and images is essential. Content must respect each platform’s style guidelines and export cleanly for direct upload.
While customer-query automation (WhatsApp/Messenger) and sales-insight modules will come later, the current architecture should leave room for those add-ons—modular code, clear APIs, and well-documented endpoints.
Acceptance criteria
1. A working prototype that handles at least 20 test products per marketplace without errors.
2. Language toggle producing grammatically correct Bangla and English; “Other languages” can route through an external translation API for now.
3. Image editor returning web-optimised PNG/JPG under 500 KB in <10 seconds on a mid-tier server.
4. Deployment guide and source code in a private repo, plus a short video walkthrough.
Technologies are flexible—Python, Node.js, or similar—with the option to plug in OpenAI or comparable LLMs for text, and an established vision API for images. Clear, maintainable code and concise documentation matter more than stack choice.