Admin backend — YukonIT (Options 1 + 2: not either/or) (2026-10-17)
Admin backend — YukonIT (Options 1 + 2: not either/or) (2026-10-17)
Young's question: the YukonIT business needs an admin backend. Option 1 = Nuxt UI Pro Dashboard; Option 2 = Medusa V2 Admin + POS + customer KB with EVE. Verified answer: these are two layers of one backend, and we need both — the question is build order.
The three consoles that actually exist
| Console | Serves | Base | Cost | Build effort |
|---|---|---|---|---|
Medusa Admin (/app) | the merchant: orders, products, inventory, refunds, POS reconciliation, draft orders from POS tablets | ships FREE inside every Medusa v2 server (MIT) | $0 | zero — enables with the storefront stack |
| Ops Dashboard (Nuxt UI Pro family, ★1,093 MIT verified) | the business operator: tenants, support tickets, agent/fleet health, onboarding, revenue view | the same Nuxt UI component family our dashboard design targets | MIT | the /dashboard design generalizes (add tenant + helpdesk panels) |
| Customer KB + EVE reception | the end-customer: self-help docs + chat | our Docus vault-pipeline + EVE (already the product) | Apache-2.0 | the BuildLibrary product = pilot 2's KB |
Recommended order (build sequence)
- Medusa Admin first — it is not optional: it's the merchant backend the storefront+POS require (draft orders from POS tablets sync here; refunds/pricing/menus are managed here). Ships with the stack we already registered (medusa-hosting-docker-study).
- EVE reception + KB second — the customer-facing layer; EVE's HIL user actions write INTO Medusa Admin's queues (order-status queries read Store API; refund requests land as draft orders awaiting admin approval).
- Ops Dashboard (Option 1) third — our internal /dashboard design already uses Nuxt UI; YukonIT's admin = same components + business panels (tenants, tickets, revenue, agent health). v1 = our fleet; v1.5 = add YukonIT business panels; later = sellable to clients (Ops Dashboard product).
Why not choose one
- Option 2 alone = merchant ops only; nothing sees agents/tenants/backups (Young's fleet blindness today is exactly that gap).
- Option 1 alone = pretty ops shell with no commerce backend underneath — can't run POS or refunds.
- Together: merchant backend (Medusa Admin) + business backend (Nuxt UI Pro) + customer self-help (KB+EVE) = the full "customer services" shape Young described, with each layer using the component/framework it was built for. Licence-clean end-to-end (MIT + MIT/Apache).
Ops note for the pilot
- All three attach under ONE nginx host for yukonit.com.au:
/storefront,/app(Medusa Admin, auth-gated, loopback-only publish),/help(KB), EVE chat widget on storefront. POS tablets hit the tenant backend via publishable key. Per-tenant pattern unchanged from the Docker design. - EVE receptionist tools verified pattern (medusa-eve-integration-study): Store API read + HIL-gated Admin writes; same keys discipline.
🎯 Tam's Read: Option 2 is the mandatory merchant layer that ships free with the stack; Option 1 is the operator layer our own dashboard design already rides — build Medusa Admin by default with the storefront, EVE+KB as the customer face, and let the ops dashboard grow out of the same Nuxt UI family we're using for the fleet — three consoles, one backend, zero licence friction.
Research
Pages in this section (25):
AgenticMail — studied and declined; and how EVE becomes customer service
Consolidates three docs into this one (agenticmail-study-2026-10-19, agenticmail-eve-study, agenticmail-collab-record). Young, 2026-10-19: "Do we needed for EVE package?" and "How do we optimise EVE to become customer service in front-end and backend for working through orders and messaging — replacing the email need."

