diff --git a/COMPLETED.md b/COMPLETED.md
index 5be3fc4..593d28b 100644
--- a/COMPLETED.md
+++ b/COMPLETED.md
@@ -60,6 +60,31 @@ Collapsed the two divergent Razor trees in `AudioPlayerBar.razor` (`@if (_isDesk
 **Notes for future work:**
 - First-render layout flash eliminated by construction (CSS media query evaluates at paint, not async subscription).
 
+### Track Card Plain-Shell Refactor
+
+**Landed 2026-06-05.**
+
+Eliminated `!important` declarations from track card CSS by replacing MudBlazor surface components with plain HTML. Implemented per `product-notes/track-card-css-architecture.md` Option A.
+
+**Razor changes (`DeepDrftShared.Client/Components/TrackCard.razor`):**
+- `MudCard` → `<div class="deepdrft-track-card-container">`
+- Fallback `MudPaper` → `<div class="deepdrft-track-card-fallback">`
+- `MudCardContent` → `<div class="deepdrft-track-card-content">`
+- `MudText`, `MudChip`, `MudFab` unchanged.
+
+**CSS changes (`DeepDrftPublic/wwwroot/styles/deepdrft-styles.css` §8):**
+- Removed four `!important` declarations from `.deepdrft-track-card-container`, `.deepdrft-track-card-fallback` base, and the dark/light theme-scoped variants.
+- Plain single-class selectors now win by cascade without `!important`; theme-scoped rules use normal specificity hierarchy.
+
+**Scope:**
+- `TrackCard` component in shared `DeepDrftShared.Client` consumed by both public site and CMS.
+- CSS in `DeepDrftPublic/wwwroot/styles/deepdrft-styles.css` (public site only).
+- Build clean: 0 errors, 0 new warnings.
+
+**Notes for future work:**
+- Plain-div shell re-enables CSS isolation as an option (a `TrackCard.razor.css` would now work against the shell divs). Section 8's public-only scoping remains convenient; isolation is optional for future polish.
+- Removes the structural mismatch of using a Material surface component (`MudCard`/`MudPaper`) solely as a layout shell. TrackCard now mirrors the construction of `NowPlayingCard` (plain divs + themed CSS).
+
 **Status:** Desktop AudioPlayerBar redesign landed on 2026-06-04.
 
 ### Desktop AudioPlayerBar — migrate to MudBlazor theme system
diff --git a/product-notes/track-card-css-architecture.md b/product-notes/track-card-css-architecture.md
new file mode 100644
index 0000000..9d51a9b
--- /dev/null
+++ b/product-notes/track-card-css-architecture.md
@@ -0,0 +1,446 @@
+# Track Card CSS Architecture — eliminate `!important` from section 8
+
+Status: completed. Author: product-designer. Date: 2026-06-05. Implementer: daniel-c-harvey. Landed: 2026-06-05.
+
+Predecessor: [`track-card-theming.md`](track-card-theming.md) (completed 2026-06-05) landed the
+navy-glass + moss-green treatment and, in doing so, introduced the two `!important` rules this
+spec exists to remove. That note is the *what it should look like*; this note is the *how the
+CSS should be structured so it looks like that without `!important`*. Nothing here changes the
+visual design — only the mechanism.
+
+---
+
+## §1. Root cause analysis — why MudCard/MudPaper fight our CSS
+
+### 1.1 The two `!important` rules in question
+
+`DeepDrftPublic/wwwroot/styles/deepdrft-styles.css` section 8:
+
+```css
+.deepdrft-track-card-container { ... background: transparent !important; }          /* line 222 */
+.deepdrft-track-card-fallback  { ... background: var(--deepdrft-navy-mid, #162437) !important; }  /* line 268 */
+.deepdrft-theme-dark  .deepdrft-track-card-fallback { ... !important; }             /* line 273 */
+.deepdrft-theme-light .deepdrft-track-card-fallback { ... !important; }             /* line 280 */
+```
+
+Four `!important` declarations, all on `background`, all on elements rendered by MudBlazor
+(`MudCard` → container, `MudPaper` → fallback).
+
+### 1.2 What MudBlazor paints and with what selector
+
+MudBlazor ships a single global `<style>` block (injected by `MudThemeProvider`, served as
+`_content/MudBlazor/MudBlazor.min.css`). The relevant rules are roughly:
+
+```css
+.mud-paper {
+    background-color: var(--mud-palette-surface);   /* the conflict */
+    color: var(--mud-palette-text-primary);
+    ...
+}
+.mud-card { /* composes .mud-paper */ }
+```
+
+`MudCard` and `MudPaper` both render a root element carrying the `.mud-paper` class, and
+`--mud-palette-surface` resolves to `DeepDrftPalettes.Surface` — `#162437` (navy-mid) in dark,
+`#FAFAF8` (off-white) in light. So **every** track card root gets a painted surface color we do
+not want, because we layer our own album-art div / fallback glass / scrim underneath the content.
+
+### 1.3 Why our scoped CSS can't reach it and `!important` was the escape hatch
+
+Two compounding facts:
+
+1. **Specificity / source-order.** `.mud-paper { background-color: ... }` is a single class
+   selector (specificity `0,1,0`). Our `.deepdrft-track-card-container { background: ... }` is
+   *also* a single class selector (`0,1,0`). When specificity ties, **source order wins** — and
+   MudBlazor's stylesheet load order relative to `deepdrft-styles.css` is not something section 8
+   controls reliably (framework CSS, theme-injected CSS, and our sheet are separate links). To
+   guarantee we win regardless of order, the author reached for `!important`. That is the smell:
+   `!important` is being used to paper over an *ordering/specificity* uncertainty, not a genuine
+   need.
+
+2. **CSS isolation cannot help here.** This is the deeper structural reason. Blazor scoped CSS
+   (`*.razor.css`) works by stamping a `b-<hash>` attribute onto elements *that Blazor renders*
+   and rewriting the scoped selector to `.foo[b-<hash>]` (specificity `0,2,0` — would beat
+   `.mud-paper`). But `MudCard`/`MudPaper` render their **own** inner HTML; Blazor never stamps
+   the scope attribute onto MudBlazor's internal elements. So a `TrackCard.razor.css` rule
+   targeting `.mud-paper` would compile to `.mud-paper[b-<hash>]` and match **nothing**. (This is
+   also why no `TrackCard.razor.css` exists today — it would be inert against the very elements
+   that need overriding.) The team correctly fell back to global rules in section 8 — but global
+   single-class rules don't beat MudBlazor without either higher specificity or `!important`.
+
+**The fundamental mismatch:** we are using MudBlazor surface components (`MudCard`, `MudPaper`)
+as *layout shells* while wanting to own their backgrounds entirely. MudBlazor's whole contract is
+"I paint the surface from the palette." We are fighting the component's purpose. Every option
+below is a different way to stop fighting it.
+
+### 1.4 What we actually need from these two elements
+
+Stripped to first principles, the container and fallback need:
+- **Container:** a `position: relative`, fixed-size, `overflow: hidden` box. A *transparent*
+  stacking context so the album-art div, fallback glass, and content scrim layer correctly.
+  No surface paint, no Material elevation (already `Elevation="0"`), no ripple.
+- **Fallback:** an absolutely-positioned full-bleed panel painted with our navy-glass (dark) or
+  navy-tint (light) treatment, with a border + `backdrop-filter`.
+
+Neither needs anything MudBlazor's surface model provides. That observation drives the
+recommendation.
+
+---
+
+## §2. Option A — Replace container + fallback with plain `<div>`
+
+Drop `MudCard` → `<div class="deepdrft-track-card-container">` and the fallback `MudPaper` →
+`<div class="deepdrft-track-card-fallback">`. Keep `MudCardContent`, `MudText`, `MudChip`,
+`MudFab` (none of these own a competing background once the card root is a plain div — and the
+text/chip/fab already work today).
+
+**Note:** `MudCardContent` technically expects a `MudCard` ancestor, but it only emits a
+`<div class="mud-card-content">` with padding — no background, no JS, no required parent context.
+It renders fine inside a plain div. If we want zero MudBlazor coupling in the shell, it can also
+become a plain `<div class="deepdrft-track-card-content">` (the class already carries all the
+real styling — padding, flex, z-index, scrim). Recommend replacing it too, for consistency.
+
+### Pros
+- **Removes the conflict at the source.** No `.mud-paper` background means no `!important`
+  anywhere. Section 8 backgrounds become plain single-class rules that win by default.
+- **Re-enables CSS isolation as an option.** With plain divs, a future `TrackCard.razor.css`
+  *would* work (Blazor stamps the scope attribute on divs it renders). We are not required to use
+  it — section 8's public-only scoping is still convenient per the predecessor note — but the door
+  reopens. This satisfies the "design the seam" instinct: we are not just fixing today, we are
+  removing a structural block.
+- **Precedent already exists in this codebase.** `NowPlayingCard.razor` is built entirely from
+  plain `<div>`s with a `.razor.css` and achieves the *exact aesthetic these cards are imitating*.
+  This option makes TrackCard structurally consistent with the component it was told to match.
+  The two should be siblings in construction, not one Material and one hand-rolled.
+- **Lighter DOM + less CSS the browser must override.** Removes `.mud-card` / `.mud-paper`
+  rulesets from applying to these nodes entirely.
+
+### Cons — what MudBlazor functionality is lost
+- **Elevation shadow:** already `Elevation="0"` on both elements (current Razor). Nothing lost —
+  the design explicitly moved to flat-glass (predecessor §3d). A plain div is the honest
+  expression of `Elevation="0"`.
+- **Ripple:** `MudCard`/`MudPaper` do **not** ripple (ripple is a `MudButton`/`MudFab`/`MudIconButton`
+  behavior). The card is not clickable as a whole — the only interactive element is the `MudFab`,
+  which stays and keeps its ripple. Nothing lost.
+- **ARIA / semantics:** `MudCard` renders a plain `<div>` with no implicit role — it adds no ARIA
+  semantics a plain div lacks. The card conveys meaning through its text content and the labeled
+  FAB, not through a container role. Nothing lost. (If anything, a future enhancement could add an
+  explicit `role`/`aria-label` to the div — easier on a plain element than fighting MudBlazor's.)
+- **Theme reactivity of the surface color:** we *want* to lose this — the whole problem is that
+  MudBlazor repaints the surface from the palette. Our section-8 rules are already theme-aware via
+  `.deepdrft-theme-dark` / `-light` wrappers, so theme switching still works.
+- **Migration cost:** ~6 lines of Razor changed. Trivial.
+
+**Assessment:** acceptable in full. There is no MudBlazor capability the track-card shell actually
+uses. This is the option that resolves the root cause rather than the symptom.
+
+---
+
+## §3. Option B — Inline `Style="background: transparent"` on MudCard/MudPaper
+
+Keep the MudBlazor components; pass an inline style to clear their background:
+
+```razor
+<MudCard Class="deepdrft-track-card-container" Elevation="0" Style="background: transparent;">
+<MudPaper Class="deepdrft-track-card-fallback" Elevation="0" Style="background: transparent;">
+```
+
+Inline styles have specificity above any selector (and above `!important` from a stylesheet, for
+non-`!important` inline declarations the inline wins over normal stylesheet rules; MudBlazor's
+`.mud-paper` background is *not* `!important`, so inline transparent beats it cleanly).
+
+But note: the fallback's *real* navy-glass paint then has to come from **somewhere**. If we set
+the MudPaper transparent inline and paint via `.deepdrft-track-card-fallback` in CSS, we are
+back to a CSS class trying to beat `.mud-paper` — except now `.mud-paper` is already transparent
+(inline), so our class wins without `!important`. That works. The container is simpler: it just
+wants transparent, which the inline gives directly.
+
+### Pros
+- No `!important`. Inline transparent neutralizes `.mud-paper`, then our classes paint normally.
+- Smaller change than Option A (no element swap), keeps MudBlazor components if there's a reason
+  to (there isn't a strong one here).
+- No risk of `MudCardContent`-without-`MudCard` edge cases.
+
+### Cons
+- **Scatters styling concern into Razor.** The whole point of section 8 is centralized,
+  theme-aware card styling. An inline `Style` on the component is a second place card appearance
+  is decided — exactly the maintenance split the team should avoid. Next person editing the card
+  background looks in section 8, doesn't find the override, and is confused.
+- **Doesn't address the root mismatch.** We're still using a surface component as a layout shell
+  and then telling it "don't be a surface." It reads as an apology for the wrong component choice.
+- **CSS isolation stays blocked** — MudBlazor still renders the inner HTML, so a `.razor.css`
+  still can't reach it. The seam stays closed.
+- **Inline `transparent` is itself a quiet "magic override."** It's `!important` by another name —
+  a higher-specificity hammer applied at the element. Marginally cleaner than `!important` in a
+  sheet, but the same category of move: beating MudBlazor by force rather than not inviting the
+  fight.
+
+**Assessment:** cleaner than `!important` strictly speaking, but trades one smell (sheet
+`!important`) for a milder one (scattered inline overrides) and leaves the structural mismatch and
+the blocked isolation seam intact. A lateral move, not a fix.
+
+---
+
+## §4. Option C — Override `--mud-palette-surface` locally on the container
+
+Scope a CSS custom-property override to the card so MudBlazor's own
+`background-color: var(--mud-palette-surface)` resolves to transparent inside the card subtree:
+
+```css
+.deepdrft-track-card-container {
+    --mud-palette-surface: transparent;
+}
+```
+
+Because `.mud-paper` reads `var(--mud-palette-surface)` and custom properties inherit, both the
+container's own paint and the nested `MudPaper` fallback would inherit the transparent value — no
+`!important` needed (we're changing the *input* to MudBlazor's rule, not overriding the rule).
+
+### Pros
+- No `!important`, no inline styles, no element swap. One declaration.
+- Works *with* MudBlazor's mechanism instead of against it — arguably the most "correct" CSS-vars
+  approach. We feed the palette variable the value we want for this subtree.
+- Centralized in section 8.
+
+### Cons
+- **Blast radius is the whole subtree.** `--mud-palette-surface` is read by *many* MudBlazor
+  components, not just paper. Any MudBlazor component nested in the card that relies on
+  `--mud-palette-surface` (now or in future) silently renders transparent. Today the card nests
+  `MudChip` and `MudFab` — both read other palette vars (`--mud-palette-primary` etc.), but a
+  future addition (a `MudMenu`, a `MudTooltip` surface, a `MudPaper` accent) would inherit the
+  transparent surface and break subtly. This is a **spooky-action-at-a-distance** override: it
+  works today but plants a trap.
+- **Fragile against MudBlazor internals.** It depends on `.mud-paper` continuing to read exactly
+  `--mud-palette-surface` for its background. If a MudBlazor version computes the surface
+  differently (e.g. a derived var, or `color-mix`), the override misses. Coupling our fix to a
+  framework implementation detail is a maintenance liability — the kind of thing that breaks on a
+  minor version bump with no compile error.
+- **Still doesn't reopen CSS isolation.** Same as B — MudBlazor renders the HTML.
+- **Obscure.** The next maintainer sees a card rendering transparent and has no obvious thread to
+  pull — the cause is a custom-property override three rules up, not a `background` declaration.
+  Low discoverability.
+
+**Assessment:** the most elegant-looking, the most dangerous. It's a clever override that depends
+on framework internals and leaks into every descendant. Reject for the trap it sets, despite the
+aesthetics.
+
+---
+
+## §5. Option D — Keep `!important` with a clear comment
+
+Leave the four `!important`s, add a comment explaining the MudBlazor specificity conflict, accept
+it as documented technical debt.
+
+### Pros
+- Zero change, zero risk. The cards render correctly today.
+- A good comment turns a mystery smell into a known, explained trade-off — which is genuinely
+  better than an *unexplained* `!important`.
+
+### Cons
+- **Doesn't meet the stated goal.** The task is explicitly to *eliminate* `!important` from
+  section 8. Documenting it is not eliminating it.
+- **`!important` is contagious.** Once the card backgrounds are `!important`, any future rule that
+  needs to legitimately override them (a "now playing" highlight state on a card, a selected
+  state, a hover treatment) must *also* be `!important`, escalating the war. The predecessor note
+  already floats a genre-chip treatment and skeleton tints — feature growth on this card is
+  expected, and each new state inherits the `!important` tax.
+- **Leaves the structural mismatch and blocked isolation seam in place** — same as B and C.
+
+**Assessment:** honest, but it's the do-nothing option dressed as a decision. Acceptable only if
+all other options were blocked, which they are not. The comment is worth keeping *as a fallback*
+if Option A is somehow rejected.
+
+---
+
+## §6. Recommendation — Option A (plain `<div>` shell)
+
+**Replace `MudCard` and the fallback `MudPaper` with plain `<div>`s. Replace `MudCardContent` with
+a plain `<div>` too (it carries no behavior). Keep `MudText`, `MudChip`, `MudFab`.** This removes
+all four `!important` declarations, preserves the glass-card aesthetic exactly (the section-8 paint
+rules are unchanged except for dropping `!important`), is the most maintainable, and aligns
+TrackCard structurally with `NowPlayingCard` — the very component it was designed to imitate.
+
+Why A over the others, in one line each:
+- **Over B (inline transparent):** A fixes the root mismatch and reopens the isolation seam; B
+  scatters overrides and leaves both problems.
+- **Over C (palette-var override):** A has no blast radius and no coupling to MudBlazor internals;
+  C plants a transparent-surface trap for every future nested component.
+- **Over D (keep `!important`):** A meets the stated goal; D doesn't.
+
+Trade-off being accepted with A: TrackCard's shell stops being a MudBlazor component. That is a
+*feature* here (the card never used MudBlazor's surface behavior), but it does mean the shell no
+longer auto-reacts to a future palette change to `Surface` — which is exactly what we want, since
+our theme-aware section-8 rules own the card's appearance. No real cost.
+
+---
+
+### 6.1 Exact Razor changes — `DeepDrftShared.Client/Components/TrackCard.razor`
+
+Replace the current file body with the following. Only the three shell elements change
+(`MudCard`→`div`, `MudPaper`→`div`, `MudCardContent`→`div`); all class hooks, `MudText`,
+`MudChip`, and `MudFab` are preserved verbatim from the current file.
+
+```razor
+<div class="deepdrft-track-card-container">
+
+    @if (!string.IsNullOrEmpty(TrackModel?.ImagePath))
+    {
+        <div class="deepdrft-track-card-bg" style="background-image: url('@TrackModel.ImagePath');">
+        </div>
+    }
+    else
+    {
+        <div class="deepdrft-track-card-fallback"></div>
+    }
+
+    <div class="deepdrft-track-card-content">
+
+        <div class="deepdrft-track-info-top">
+            <MudText Typo="Typo.subtitle1"
+                     Class="deepdrft-track-title text-truncate mb-1">
+                @TrackModel?.TrackName
+            </MudText>
+
+            <MudText Typo="Typo.caption"
+                     Class="deepdrft-track-artist text-truncate mb-2">
+                @TrackModel?.Artist
+            </MudText>
+        </div>
+
+        <div class="deepdrft-track-info-middle">
+            @if (!string.IsNullOrEmpty(TrackModel?.Album))
+            {
+                <MudText Typo="Typo.caption"
+                         Class="deepdrft-track-meta text-truncate">
+                    @TrackModel.Album
+                </MudText>
+            }
+
+            @if (!string.IsNullOrEmpty(TrackModel?.Genre))
+            {
+                <MudChip T="string"
+                         Size="Size.Small"
+                         Variant="Variant.Filled"
+                         Color="Color.Primary"
+                         Class="deepdrft-genre-chip">
+                    @TrackModel.Genre
+                </MudChip>
+            }
+        </div>
+
+        <div class="deepdrft-track-info-bottom">
+            @if (TrackModel?.ReleaseDate.HasValue == true)
+            {
+                <MudText Typo="Typo.caption"
+                         Class="deepdrft-track-meta">
+                    @TrackModel.ReleaseDate.Value.Year
+                </MudText>
+            }
+            else
+            {
+                <div></div>
+            }
+
+            <MudFab Color="Color.Primary"
+                    Size="Size.Medium"
+                    StartIcon="@PlayPauseIcon"
+                    OnClick="@PlayClick"/>
+        </div>
+
+    </div>
+
+</div>
+```
+
+Changes, precisely:
+- Line 1–2: `<MudCard Class="deepdrft-track-card-container" Elevation="0">` →
+  `<div class="deepdrft-track-card-container">`. (Drop `Elevation="0"` — meaningless on a div; the
+  flat look is the default.)
+- Line 11–13: `<MudPaper Class="deepdrft-track-card-fallback" Elevation="0"></MudPaper>` →
+  `<div class="deepdrft-track-card-fallback"></div>`.
+- Line 16: `<MudCardContent Class="deepdrft-track-card-content">` →
+  `<div class="deepdrft-track-card-content">`.
+- Line 72 (`</MudCard>`) → `</div>`; the `</MudCardContent>` close (line 70) → `</div>`.
+
+No change to `TrackCard.razor.cs` — it references no MudBlazor card types (only `Icons`,
+`TrackDto`, `EventCallback`). The `using MudBlazor;` stays (still used by `Icons.Material`).
+
+### 6.2 Exact CSS changes — `DeepDrftPublic/wwwroot/styles/deepdrft-styles.css` section 8
+
+Remove `!important` from all four declarations. No other change — the colors, layout, and
+theme-aware scoping all stay exactly as they are.
+
+**Line 216–223** — container:
+```css
+.deepdrft-track-card-container {
+    width: 250px;
+    height: 250px;
+    min-width: 250px;
+    position: relative;
+    overflow: hidden;
+    background: transparent;        /* was: transparent !important */
+}
+```
+
+**Line 262–269** — fallback base:
+```css
+.deepdrft-track-card-fallback {
+    position: absolute;
+    top: 0;
+    left: 0;
+    width: 100%;
+    height: 100%;
+    background: var(--deepdrft-navy-mid, #162437);   /* was: ... !important */
+}
+```
+
+**Line 272–276** — fallback dark:
+```css
+.deepdrft-theme-dark .deepdrft-track-card-fallback {
+    background: color-mix(in srgb, var(--deepdrft-navy) 55%, transparent);   /* was: ... !important */
+    border: 1px solid rgba(250, 250, 248, 0.12);
+    backdrop-filter: blur(8px);
+}
+```
+
+**Line 279–282** — fallback light:
+```css
+.deepdrft-theme-light .deepdrft-track-card-fallback {
+    background: color-mix(in srgb, var(--deepdrft-navy) 8%, var(--deepdrft-white));   /* was: ... !important */
+    border: 1px solid var(--deepdrft-border);
+}
+```
+
+After this change, the four backgrounds are plain single-class (and class-under-theme-wrapper)
+rules painting plain `<div>`s. Nothing competes; nothing needs `!important`. The
+`.deepdrft-theme-dark .deepdrft-track-card-fallback` selector is `0,2,0` and the base is `0,1,0`,
+so the theme override correctly beats the base — the normal cascade, working as intended.
+
+### 6.3 Verification checklist for the implementer
+- Dark public site: card with album art shows art + scrim + green/off-white text; card without
+  art shows navy-glass fallback. No grey/teal surface flash on hydration (the base
+  `--deepdrft-navy-mid` fallback rule still covers the no-wrapper-class window).
+- Light public site: fallback shows the navy-tint-on-white panel; text legible.
+- Toggle dark/light: card repaints correctly (theme-wrapper-scoped rules still drive it).
+- Confirm no other component references `.deepdrft-track-card-container` / `-fallback` expecting a
+  `.mud-paper` ancestor (grep — `TracksGallery.razor` wraps cards but only positions them via
+  `.deepdrft-track-gallery-item-center`).
+- Confirm the rendered card root is now a plain `<div>` (devtools) with **no** `.mud-paper` /
+  `.mud-card` class, and that no `background ... !important` remains anywhere in section 8.
+
+### 6.4 What this unblocks (not in scope, noted for the seam)
+With a plain-div shell, the deferred polish items from the predecessor note become cleaner to
+implement and could, if desired, migrate to a real `TrackCard.razor.css` (now functional):
+genre-chip distinct treatment (predecessor §3a), skeleton tints (§3c), card hover/selected/now-
+playing states. None are required by this spec; flagged so a future pass knows the seam is open.
+
+---
+
+## §7. Fallback position
+
+If Option A is rejected for a reason not surfaced here (e.g. a policy that shell components must
+remain MudBlazor for consistency with other cards), the next-best is **Option D with comments** —
+not B or C. B scatters the override and C couples to framework internals; a *documented*
+`!important` is more honest and more maintainable than either clever hack. But A has no real cost
+identified, so this fallback should not be needed.
+```
\ No newline at end of file