SNAPLIST

Inspiration

500 million small manufacturers worldwide make incredible products but remain invisible on global ecommerce. Getting on Amazon requires studio photos, SEO titles, and formatted listings. Hiring someone costs $5,000 and takes 3 weeks. We built SNAPLIST to eliminate that barrier completely in 60 seconds.

What it does

SNAPLIST transforms any product photo into a complete marketplace-ready listing instantly. Upload a photo, describe your product in one sentence, and get:

• AI background removal via remove.bg API
• SEO-optimized title following Amazon's exact format requirements
• 5 specific bullet points with measurable product details — e.g. "280gsm cotton-polyester blend resists shrinking after 50+ washes"
• 200 word marketplace-specific description
• 10 targeted search keywords
• Marketplace readiness score
• Export as Amazon CSV, Shopify CSV, or ZIP
• Multi-turn refinement chat to update any field

How we built it

Built entirely on MeDo's full-stack AI platform using multi-turn conversational generation.

Tech Stack:

• MeDo — full-stack app generation
• Supabase — PostgreSQL database and Edge Functions
• remove.bg API — AI background removal for product photos
• Groq API with Llama 3.3 70B — AI listing generation with expert Amazon copywriting system prompts
• React + Tailwind CSS — frontend UI
• Email authentication — persistent user accounts
• CSV export engine — Amazon Seller Central and Shopify exact column formats

We used MeDo's conversational interface to architect everything from database schema to React components to Supabase Edge Functions. The entire pipeline — from photo upload to CSV export — was built through MeDo's multi-turn generation without writing a single line of code manually.

Challenges we ran into

Challenge 1 — AI listing quality: Early versions generated generic content like "Made from quality materials" — useless for real sellers. We solved this by engineering a specialized Amazon copywriting system prompt that bans generic phrases and requires measurable details in every bullet. Result: listings like "280gsm cotton-polyester blend resists shrinking after 50+ washes" instead of "made from quality fabric."

Challenge 2 — Image processing: MeDo's native image handling couldn't process user-uploaded photos for background removal. We solved this by integrating remove.bg API directly into Supabase Edge Functions, then adding smart contrast detection to handle edge cases gracefully with helpful user guidance.

Challenge 3 — Marketplace formatting: Amazon and Shopify have completely different listing formats. We built a marketplace-specific formatting engine that generates different content structure, tone, and field lengths for each platform automatically.

Accomplishments that we're proud of

• Complete end-to-end catalog generation pipeline built in 4 days using MeDo
• AI listing quality that matches professional Amazon copywriters — specific, measurable, conversion-focused bullets
• Processing animation with rotating e-commerce education facts that makes the 30-second wait feel engaging
• Marketplace-specific formatting for Amazon, Shopify, Etsy, Meesho, and Tokopedia
• Vague description detection that guides users toward better inputs automatically

What we learned

MeDo's multi-turn conversation approach is genuinely powerful for rapid full-stack development. Iterating on backend Edge Functions, frontend components, and database schema through conversation dramatically accelerated our development cycle.

We also learned that prompt engineering for domain-specific AI tasks requires deep domain knowledge. Generic prompts produce generic output. Specialized system prompts that encode professional expertise — like Amazon listing copywriting rules — produce professional results.

What's next

• Bulk processing for 100+ SKUs simultaneously
• Direct Amazon Seller Central API integration for one-click publishing
• Regional marketplace support in local languages — Hindi for Meesho, Bahasa for Tokopedia
• AI vision integration to analyze product images directly without requiring text description

Built With

• groq
• javascript
• llama-3.3-70b
• medo
• react
• remove.bg-api
• supabase
• tailwind-css