github-amboras

GitHub README for xaid-vfx/amboras. Homepage: amboras.okzaid.com. Description: "The Shopify of the AI Era — brand brief to live store in under 5 minutes."

Key extractions

• Thesis: founder types plain-English brand brief → fully generated Shopify-backed storefront in under 5 minutes. No design, no dev.
• Flow: /api/generate opens SSE stream → GPT-4o-mini via OpenRouter generates typed StoreConfig via tool-call → Shopify Admin REST creates real products/variants → persist (file locally, Upstash Redis in prod).
• Key decision: structured tool output over free-form JSON — no parsing risk, no partial failures. Single AI call produces typed config.
• Key decision: SSE over WebSockets — works with Vercel's streaming response model, no persistent connection.
• Key decision: Shopify as commerce layer — real inventory, real payments, real checkout. Amboras only handles storefront + AI.
• Key decision: file storage locally, Redis in prod, same interface — zero-config dev.
• Key decision: client-side cart state — no cart DB; Shopify checkout built from variant IDs at checkout time.
• Data model: StoreConfig has brand, hero, products, testimonials, stats, about, faq, seo. AI editor produces patches against this same schema.

Patterns

• SSE for AI streaming: generation + edit both use SSE → sse-for-ai-streaming.
• Structured output via tool calls: ai-model-choice.
• Compress mechanical labor: brief → full store. compress-mechanical-labor.