diff --git a/README.md b/README.md
index 37eba6d..9584432 100644
--- a/README.md
+++ b/README.md
@@ -299,6 +299,7 @@ Every theme, font, layout, and behaviour knob lives in `preferences`. Override a
 |---|---|---|
 | `font` | `"Lato"` | Primary font family. Must be installed. |
 | `bodySize` | `10pt` | Base text size. Sub-elements scale via em-multipliers. |
+| `density` | `"comfortable"` | Single knob that scales every vertical-spacing em-token uniformly — block above/below, `v()` gaps, `par` leading/spacing, `list` spacing. `"compact"` (× 0.85) tightens for a one-page fit, `"comfortable"` (× 1.0) is the historical default, `"spacious"` (× 1.15) opens it up for print presentation. Text sizes, icon dimensions, and rating-dot geometry are unaffected — use `bodySize` for those. Unknown values panic. |
 | `paper` | `"a4"` | Paper size string passed to Typst's `set page(paper: ...)`. `"a4"`, `"us-letter"`, `"a5"`, `"us-legal"`, and the rest of [Typst's named papers](https://typst.app/docs/reference/layout/page/#parameters-paper). |
 | `margin` | `(x: 0.9cm, y: 1.5cm)` | Page margins. Anything `set page(margin: ...)` accepts. |
 | `accent` | `palettes.teal` | Theme colour for headings, accent rules, tags, dots. Use a built-in preset — `palettes.{teal,navy,crimson,forest,plum,charcoal}` — or any `rgb(...)` value. |
diff --git a/examples/tests/density.pdf b/examples/tests/density.pdf
new file mode 100644
index 0000000..63fd6d7
Binary files /dev/null and b/examples/tests/density.pdf differ
diff --git a/internal/header.typ b/internal/header.typ
index bcd5d75..1c9dcc8 100644
--- a/internal/header.typ
+++ b/internal/header.typ
@@ -5,7 +5,7 @@
 // `_format_location` / `_url_encode` helpers that feed the maps deep
 // link (single-caller helpers, kept adjacent to their consumer).
 
-#import "state.typ": _body_size_state, _accent_state, _emphasis_colour
+#import "state.typ": _body_size_state, _accent_state, _spacing_scale_state, _emphasis_colour
 #import "presets.typ": maps-providers
 #import "icons.typ": icon, _profile_networks, _network_aliases
 
@@ -146,11 +146,12 @@
   context {
     let body-size = _body_size_state.get()
     let accent = _accent_state.get()
+    let scale = _spacing_scale_state.get()
 
     let header-text = align(text-align, {
       block(
         spacing: 0pt,
-        below: 1.2 * body-size,
+        below: 1.2 * scale * body-size,
         text(
           2.5 * body-size,
           fill: accent,
@@ -162,7 +163,7 @@
       if "label" in basics and basics.label != none {
         block(
           spacing: 0pt,
-          below: 0.8 * body-size,
+          below: 0.8 * scale * body-size,
           text(1.2 * body-size, fill: _emphasis_colour, weight: "bold", basics.label),
         )
       }
@@ -308,7 +309,7 @@
         // the document width, regardless of how the text above /
         // below is aligned.
         let centred-photo = block(
-          spacing: 0.8 * body-size,
+          spacing: 0.8 * scale * body-size,
           width: 100%,
           align(center, photo),
         )
@@ -335,7 +336,8 @@
   let summary = basics.at("summary", default: none)
   if summary == none or summary == "" or summary == [] { return }
   let body-size = _body_size_state.get()
-  v(0.8 * body-size)
+  let scale = _spacing_scale_state.get()
+  v(0.8 * scale * body-size)
   par(summary)
-  v(0.4 * body-size)
+  v(0.4 * scale * body-size)
 }
diff --git a/internal/layout.typ b/internal/layout.typ
index dde8b27..cd31888 100644
--- a/internal/layout.typ
+++ b/internal/layout.typ
@@ -90,6 +90,12 @@
   // Every spacing token is an em-multiplier of this, so changing one
   // knob scales the whole document proportionally.
   bodySize: 10pt,
+  // Scales every spacing em-token uniformly (see `_density_scales`
+  // in `internal/state.typ`). `"comfortable"` (1.0×) preserves the
+  // historical layout byte-for-byte; `"compact"` (0.85×) tightens
+  // for one-page CVs; `"spacious"` (1.15×) opens it up for print
+  // presentation. Text sizes are unaffected.
+  density: "comfortable",
   paper: "a4",
   margin: (x: 0.9cm, y: 1.5cm),
   // `palettes.teal` — see the `palettes` dict for the curated set
diff --git a/internal/primitives.typ b/internal/primitives.typ
index e917a64..492f4a1 100644
--- a/internal/primitives.typ
+++ b/internal/primitives.typ
@@ -3,7 +3,7 @@
 // from `lib.typ` for callers composing custom layouts; the leading-
 // underscore helpers are used internally by the section renderers.
 
-#import "state.typ": _body_size_state, _accent_state, _body_colour, _divider_colour
+#import "state.typ": _body_size_state, _accent_state, _spacing_scale_state, _body_colour, _divider_colour
 #import "icons.typ": icon
 
 // Bold accent-coloured line — designed for the company / institution
@@ -11,9 +11,10 @@
 #let name(body) = context {
   let body-size = _body_size_state.get()
   let accent = _accent_state.get()
+  let scale = _spacing_scale_state.get()
   block(
     above: 0pt,
-    below: 0.6 * body-size,
+    below: 0.6 * scale * body-size,
     text(weight: "bold", fill: accent, body),
   )
 }
@@ -23,9 +24,10 @@
 #let term(period, location: none) = context {
   if period == none and location == none { return }
   let body-size = _body_size_state.get()
+  let scale = _spacing_scale_state.get()
   block(
     above: 0pt,
-    below: 0.8 * body-size,
+    below: 0.8 * scale * body-size,
     inset: (left: 0.3 * body-size),
     text(0.9 * body-size, {
       if period != none {
@@ -77,12 +79,13 @@
 #let divider() = context {
   let body-size = _body_size_state.get()
   if here().position().y < 2cm { return }
-  v(0.3 * body-size)
+  let scale = _spacing_scale_state.get()
+  v(0.3 * scale * body-size)
   line(
     length: 100%,
     stroke: (paint: _divider_colour, thickness: 0.6pt, dash: "dashed"),
   )
-  v(0.3 * body-size)
+  v(0.3 * scale * body-size)
 }
 
 // Like `divider()` but with a leading label that sits slightly indented
@@ -94,8 +97,9 @@
 // the parent section title.
 #let _labelled_divider(label) = context {
   let body-size = _body_size_state.get()
+  let scale = _spacing_scale_state.get()
   let stroke = (paint: _divider_colour, thickness: 0.6pt, dash: "dashed")
-  v(0.3 * body-size)
+  v(0.3 * scale * body-size)
   pad(left: 0.6 * body-size, grid(
     columns: (1.3em, auto, 1fr),
     column-gutter: 0.5 * body-size,
@@ -108,7 +112,7 @@
     ),
     line(length: 100%, stroke: stroke),
   ))
-  v(0.3 * body-size)
+  v(0.3 * scale * body-size)
 }
 
 // Interleaves `divider()` between items; the trailing one is suppressed
diff --git a/internal/state.typ b/internal/state.typ
index 1e148e7..1029111 100644
--- a/internal/state.typ
+++ b/internal/state.typ
@@ -11,6 +11,23 @@
 #let _body_size_state = state("alta-body-size", 10pt)
 #let _accent_state = state("alta-accent", palettes.teal)
 #let _max_rating_state = state("alta-max-rating", 5)
+// Multiplier applied to every spacing em-token (block above/below,
+// `v()`, par.spacing/leading, list.spacing). Driven by
+// `preferences.density`; defaults to 1.0 so "comfortable" reproduces
+// the historical layout byte-for-byte.
+#let _spacing_scale_state = state("alta-spacing-scale", 1.0)
+
+// `preferences.density` → multiplier on every spacing em-token.
+// 0.85 / 1.0 / 1.15 keeps the three presets visibly distinct without
+// either crushing lines together or pushing the one-page CV onto two.
+// Text sizes, icon dimensions, and rating-dot geometry are
+// deliberately left alone — density is purely vertical whitespace,
+// so font-size scaling stays the job of `bodySize`.
+#let _density_scales = (
+  compact: 0.85,
+  comfortable: 1.0,
+  spacious: 1.15,
+)
 
 // Accent is configurable via `alta(preferences: (accent: ...))`; the
 // rest are opinionated visual constants.
diff --git a/lib.typ b/lib.typ
index 23d178e..d1dfed4 100644
--- a/lib.typ
+++ b/lib.typ
@@ -19,7 +19,7 @@
 // independent of text size.
 
 #import "internal/presets.typ": palettes, maps-providers
-#import "internal/state.typ": _body_size_state, _accent_state, _max_rating_state, _body_colour, _emphasis_colour
+#import "internal/state.typ": _body_size_state, _accent_state, _max_rating_state, _spacing_scale_state, _density_scales, _body_colour, _emphasis_colour
 #import "internal/defaults.typ": _default_labels
 #import "internal/validation.typ": _strict_merge, _check_bool
 #import "internal/text.typ": _present, styled-link
@@ -140,9 +140,19 @@
   }
   let accent = preferences.accent
   let body-size = preferences.bodySize
+  let density = preferences.density
+  if density not in _density_scales {
+    let quote(k) = "\"" + k + "\""
+    panic(
+      "density must be one of " + _density_scales.keys().map(quote).join(", ")
+        + ", got: " + repr(density),
+    )
+  }
+  let scale = _density_scales.at(density)
   _accent_state.update(accent)
   _body_size_state.update(body-size)
   _max_rating_state.update(max-rating)
+  _spacing_scale_state.update(scale)
 
   // PDF metadata is sourced from `basics` (title, author, description)
   // and the JSON Resume `meta` block (date, keywords). Each optional
@@ -191,12 +201,12 @@
     margin: preferences.margin,
     footer: resolved-footer,
   )
-  set par(leading: 0.55em, spacing: 0.7em)
+  set par(leading: 0.55 * scale * 1em, spacing: 0.7 * scale * 1em)
   set list(
     marker: text(0.85em, "•"),
     indent: 0pt,
     body-indent: 0.4 * body-size,
-    spacing: 0.55em,
+    spacing: 0.55 * scale * 1em,
   )
 
   // Heading levels map to semantic CV roles:
@@ -204,21 +214,21 @@
   //   ===  role / qualification line
   //   ==== sub-grouping (publication type)
   show heading.where(level: 2): it => block(sticky: true)[
-    #v(0.6 * body-size)
+    #v(0.6 * scale * body-size)
     #text(1.7 * body-size, fill: accent, weight: "bold", upper(it.body))
-    #v(-0.7 * body-size)
+    #v(-0.7 * scale * body-size)
     #line(length: 100%, stroke: 2pt + accent)
-    #v(0.2 * body-size)
+    #v(0.2 * scale * body-size)
   ]
   show heading.where(level: 3): it => block(
-    above: 1.0 * body-size,
-    below: 0.8 * body-size,
+    above: 1.0 * scale * body-size,
+    below: 0.8 * scale * body-size,
     sticky: true,
     text(1.2 * body-size, fill: _emphasis_colour, weight: "regular", it.body),
   )
   show heading.where(level: 4): it => block(
-    above: 0.6 * body-size,
-    below: 0.6 * body-size,
+    above: 0.6 * scale * body-size,
+    below: 0.6 * scale * body-size,
     sticky: true,
     text(1.2 * body-size, fill: _emphasis_colour, weight: "bold", it.body),
   )
diff --git a/sections/projects.typ b/sections/projects.typ
index 1a9457f..dfc51e1 100644
--- a/sections/projects.typ
+++ b/sections/projects.typ
@@ -7,6 +7,7 @@
 #import "../internal/text.typ": _present, styled-link
 #import "../internal/primitives.typ": term, _join_with_dividers, _tag_row
 #import "../internal/dates.typ": _format_date_range
+#import "../internal/state.typ": _spacing_scale_state, _body_size_state
 
 #let _projects(entries, labels, prefs) = {
   let valid = entries.filter(p => _present(p.at("name", default: none)))
@@ -19,11 +20,14 @@
     if _present(description) {
       // Softer than `name()` (which is bold + accent) so the
       // description doesn't compete visually with a linked title.
-      // Wrapped in a block so the gap to the term row below matches
-      // the institution-line → term spacing in `_experience` /
-      // `_education`; using a bare `linebreak()` here leaves no
-      // paragraph spacing.
-      block(below: 0.6em, emph(description))
+      // Mirrors `name()`'s `below` so the description → term gap
+      // matches the institution-line → term gap in `_experience` /
+      // `_education`; literal `0.6em` would skip the density scale
+      // and break that contract under non-default density.
+      context block(
+        below: 0.6 * _spacing_scale_state.get() * _body_size_state.get(),
+        emph(description),
+      )
     }
     term(_format_date_range(project, prefs, labels))
     for bullet in project.at("highlights", default: ()) [- #bullet]
diff --git a/sections/skills.typ b/sections/skills.typ
index ac2fc77..02d33ca 100644
--- a/sections/skills.typ
+++ b/sections/skills.typ
@@ -4,7 +4,7 @@
 // `_name_keywords_section`; the two public renderers differ only in
 // which `labels.*` heading they pass.
 
-#import "../internal/state.typ": _body_size_state
+#import "../internal/state.typ": _body_size_state, _spacing_scale_state
 #import "../internal/primitives.typ": tag, _tag_row
 
 // `text("-")` (not `[-]`) — markup-bracketed `-` parses as a list-item
@@ -20,7 +20,8 @@
   if visible.len() == 0 { return }
   context {
     let body-size = _body_size_state.get()
-    let row-gap = 0.7 * body-size
+    let scale = _spacing_scale_state.get()
+    let row-gap = 0.7 * scale * body-size
     [== #heading]
     for group in visible {
       block(above: 0pt, below: row-gap, par(hanging-indent: 1em, leading: row-gap, {
diff --git a/tests/density.typ b/tests/density.typ
new file mode 100644
index 0000000..bea53a1
--- /dev/null
+++ b/tests/density.typ
@@ -0,0 +1,64 @@
+// `preferences.density` scales every spacing em-token uniformly so a
+// single knob trades vertical breathing room for fit-on-one-page.
+// Three pages exercise each preset against the same data; visual diff
+// against the default ("comfortable") confirms the multiplier reaches
+// section headings, divider rules, par.spacing, the in-section block
+// above/below tokens, and the header / summary gaps.
+
+#import "../lib.typ": alta
+
+#let cv = (
+  basics: (
+    name: "Jane Doe",
+    label: "Senior Software Engineer",
+    summary: [Backend engineer; the summary block exercises the
+      pre/post `v()` gaps that flank it.],
+    email: "jane@example.com",
+    phone: "+353 1 555 0100",
+    location: "Dublin, Ireland",
+  ),
+  work: (
+    (
+      name: "Acme Corp",
+      position: "Senior Software Engineer",
+      location: "Dublin, Ireland",
+      startDate: "Jan 2022",
+      highlights: ([Led the migration.], [Designed the platform.]),
+    ),
+    (
+      name: "Foo Ltd",
+      position: "Software Engineer",
+      startDate: "Jan 2018",
+      endDate: "Dec 2021",
+      highlights: ([Shipped the thing.],),
+    ),
+  ),
+  skills: (
+    (name: "Languages", keywords: ("Scala", "Python")),
+    (name: "Infra",     keywords: ("Kafka", "AWS", "Kubernetes")),
+  ),
+  languages: (
+    (language: "English", fluency: "Native"),
+    (language: "Irish",   fluency: "Professional Working"),
+  ),
+  education: (
+    (institution: "Example U", studyType: "B.Sc.", endDate: "2017"),
+  ),
+  certificates: (
+    (name: "CKA", issuer: "CNCF"),
+  ),
+)
+
+// 1. Compact — em spacing tokens × 0.85. Tighter than the default.
+#alta(cv, preferences: (density: "compact"))
+
+#pagebreak()
+
+// 2. Comfortable — em spacing tokens × 1.0. The default; explicit
+//    here so the regression is anchored even if the default changes.
+#alta(cv, preferences: (density: "comfortable"))
+
+#pagebreak()
+
+// 3. Spacious — em spacing tokens × 1.15. Roomier than the default.
+#alta(cv, preferences: (density: "spacious"))