Can I vibe code SleekFlow?

Build a unified omnichannel support agent application using Next.js (App Router), Tailwind CSS, Supabase, and WebSockets.

1. DATA MODEL:
- Contacts: id, phone_number, channel (whatsapp, telegram, web), external_id, created_at.
- Conversations: id, contact_id, assigned_agent_id, status (open, pending, closed), last_message_at.
- Messages: id, conversation_id, sender_type (agent, customer, bot), content, metadata (JSON), status (sent, delivered, read), created_at.
- Templates: id, name, channel, body_text, meta_template_id, status (approved, rejected).

2. CORE FUNCTIONALITY:
- Unified Multi-Agent Inbox: Split view with conversation list on left, active chat window in middle, and Shopify customer context sidebar on right.
- Real-time Updates: Use Supabase Realtime/WebSockets to push incoming/outgoing messages instantly to all agent dashboards.
- Meta WhatsApp Integration: Inbound webhook endpoint handling text/media payloads, signature verification, and outbound sending via Meta Cloud API.
- Automated AI Triage: When a new message arrives, route to an OpenAI GPT-4o mini pipeline. If customer asks for order status, query Shopify Admin API and return formatted tracking response. If query is complex, set conversation status to 'open' and assign to human agent pool.

3. EDGE CASES & FAILURE MODES:
- Message Deduplication: Meta webhooks can send duplicate notifications; enforce idempotency keys based on Meta message ID.
- Out-of-window Messages: Automatically block non-template outbound WhatsApp messages if the 24-hour customer window has elapsed, prompting agent to select an approved template.
- Rate Limits: Queue outgoing broadcast messages using Redis/BullMQ to strictly observe Meta rate limits (80 messages/sec).

4. OUT OF SCOPE:
- Native WeChat SDK integration, native mobile apps, automated video processing.

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