Can I vibe code Chatfuel?

Build a complete AI-powered e-commerce customer support chatbot system using Next.js, OpenAI API, and the Shopify Admin API.

1. DATA MODEL & BACKEND
- Implement a PostgreSQL database with Prisma ORM for storing Conversations, Messages, Customers, and Order Cache.
- Create an integration layer connecting to the Shopify Admin API to fetch catalog items, stock status, policy pages, and customer order history using secure order lookup endpoints.

2. AI SUPPORT ENGINE
- Integrate OpenAI API (GPT-4o-mini) using the Vercel AI SDK with function calling enabled.
- Define tools: `search_products(query)`, `get_order_status(order_number, email)`, and `get_store_policy(topic)`.
- Restrict LLM system instructions: maintain a strict brand tone, never grant unauthorized custom discounts, refuse out-of-context requests, and require email verification before returning tracking links.

3. WEB CHAT WIDGET & INBOX
- Build a lightweight React floating chat widget component that streams responses from the Next.js API route.
- Build a basic internal admin dashboard route `/admin/conversations` where human agents can view chat logs and click a 'Take Over' button to toggle automated AI responses off for a specific session.

4. EDGE CASES & SAFETY
- Handle Shopify API rate limits with exponential backoff.
- Implement fallback messages when the LLM returns errors or confidence is low.
- Sanitize user input to prevent prompt injection attacks aimed at revealing system prompts or modifying order states.

OUT OF SCOPE: Native Meta Developer App verification workflows and visual drag-and-drop node flow builders.

$ each button prefixes agent-specific run instructions · build your own product, never copy proprietary code, trademarks or designs