Yukon IT admin backend — Option 1 vs Option 2, priced
Yukon IT admin backend — Option 1 vs Option 2, priced
Young, 2026-10-17: Yukon IT (public website selling services) needs a backend admin, EVE reception, storefront + POS, landing + domain + O365.
The headline finding: both options cost $0 in licence. The choice is not a purchase decision — it is a scope decision. Everything below is verified today.
1. Licence costs — both zero, and one changed recently
| Option | Licence | Cost | Verified |
|---|---|---|---|
| Option 1 — Nuxt UI Dashboard | MIT | $0 | ui.nuxt.com: "Nuxt UI and Nuxt UI Pro are now unified into a single, fully open-source and free library" — v4.12.0, "Published under MIT License" |
| Option 2 — Medusa V2 Admin + POS | MIT core (open-core) | $0 self-hosted | github.com/medusajs/medusa: "The core is licensed under the MIT License"; "free and unlimited users for the Medusa Admin Dashboard" |
Important correction to a common assumption: Nuxt UI Pro used to cost
$249 / $499 / $999 (still visible on the old ui3.nuxt.com page). v4 made it
free and MIT. Anyone pricing Option 1 off the old page would be overstating it by
$249+. The only thing still paid is the Figma Kit ($149) — and we don't need it.
And we already have it installed: @nuxt/ui is present in docs-site's
node_modules (transitive via Docus). So Option 1's dependency cost is zero work
to add — it is already on the box.
2. So what actually differs? Scope, not price
| Option 1 — Nuxt UI Dashboard | Option 2 — Medusa Admin + POS | |
|---|---|---|
| What it is | A generic admin UI kit (multi-column layout, charts, date pickers) | A finished commerce back-office — products, orders, customers, inventory, POS |
| What it covers | Anything we build | Only commerce |
| What it does NOT cover | Anything — it's a shell, every screen is ours to build | Domains, O365, EVE, KB, billing — Medusa has no concept of these |
| Build effort | Every screen from scratch | Products/orders/customers already built and maintained |
| Maintained by | Us | Medusa (upstream) |
| Customisation limit | Unlimited | ⚠️ "You can't customize the admin dashboard's layout, design, or content of existing pages (aside from injecting widgets)" |
The Medusa limitation matters. Its admin is a finished product we can extend with widgets but not restyle. For Yukon IT that is fine for commerce and wrong for everything else.
3. The recommendation: both, layered — not either/or
Yukon IT's list splits cleanly down the middle:
Commerce (Medusa's world) Services (nobody's world — ours)
storefront landing page
POS domain name
products / orders / customers O365 email
inventory EVE reception console
→ Medusa Admin handles it KB management
(free, built, upstream-maintained) billing / plan
→ Nuxt UI Dashboard builds it
Layer 1 — Medusa Admin + POS (Option 2) for commerce ops. Do not rebuild products/orders/customers. It ships with Medusa v2, it is free, and it is maintained upstream. Our POS fork (Agilo) already speaks to it.
Layer 2 — Nuxt UI Dashboard (Option 1) as the operator portal. This is the "business backend" Young means: one login showing the tenant's whole estate — EVE reception, KB, storefront status, domain, O365, billing. It links into Medusa Admin for commerce rather than duplicating it.
Why not Medusa-Admin-only: it cannot show a domain renewal, an O365 seat count, or an EVE conversation. Yukon IT is selling services, not only products — the admin must cover what they sell.
Why not Nuxt-UI-only: we would rebuild a commerce back-office that already exists, free, upstream. That is the expensive mistake.
4. Cost of the recommended build
| Line | Cost |
|---|---|
| Licences | $0 (both MIT) |
| Nuxt UI dependency | $0 — already in the tree |
| Medusa Admin | $0 — ships with Medusa |
| Build labour | the only real cost — the portal screens |
| Hosting | fits the measured prod envelope (4 tenants; portal ≈ 0.2–0.5 GB) |
Nothing here needs a purchase approval. It needs a build decision.
5. One thing to keep separate (a real trap)
There is already a dashboard plan: fleet-dashboard-plan.md (+ Sol's annex).
That is OUR ops dashboard — fleet health, agents, costs. Yukon IT's admin is a
different product for a different audience (their staff, their data).
Conflating them would put our fleet internals behind a customer's login. They share the Nuxt UI stack; they must not share a deployment. Worth stating in the register so nobody merges them later.
6. Decisions for Young
- Layered build approved? (Medusa Admin for commerce + Nuxt UI portal for services)
- Portal scope for pilot — which screens ship first: EVE reception, KB, or billing?
- Customer-facing or internal-first? Does Yukon IT's staff log in, or do we operate it for them initially?
Honest limits
- Medusa Enterprise Edition (RBAC, SSO, audit logs) needs a commercial agreement — we are not using it. Core MIT is enough for the pilot; flagging so nobody assumes RBAC is included.
- I priced licences and hosting; I did not estimate build hours — that is Sol's call, and it is the number that actually matters here.
- The "4 tenants" envelope is RAM-based; a portal per tenant is small but not zero.