diff --git a/caption.go b/caption.go
new file mode 100644
index 00000000..53d75f3a
--- /dev/null
+++ b/caption.go
@@ -0,0 +1,453 @@
+package main
+
+import (
+	"fmt"
+	"math"
+	"os"
+	"path/filepath"
+	"runtime"
+	"strings"
+	"text/template"
+)
+
+// OverlayEvent represents a single text overlay at a specific point in the recording.
+type OverlayEvent struct {
+	StartMs    int64
+	DurationMs int64
+	Text       string
+}
+
+// OverlayOptions holds configuration for text overlays.
+type OverlayOptions struct {
+	Font           string
+	FontSize       int
+	Alignment      CaptionAlignment
+	FontColor      string // #RRGGBB hex
+	BoxColor       string // #RRGGBB hex
+	BoxOpacity     float64
+	BoxPadding     int
+	MarginLeft     int
+	MarginRight    int
+	MarginVertical int
+}
+
+const defaultOverlayDurationMs = 3000
+
+// DefaultOverlayOptions returns overlay options with sensible defaults.
+func DefaultOverlayOptions() OverlayOptions {
+	font := defaultOSFont()
+	return OverlayOptions{
+		Font:           font,
+		FontSize:       22,
+		Alignment:      AlignTopCenter,
+		FontColor:      "#FFFFFF",
+		BoxColor:       "#000000",
+		BoxOpacity:     0.5,
+		BoxPadding:     10,
+		MarginLeft:     20,
+		MarginRight:    20,
+		MarginVertical: 20,
+	}
+}
+
+// CaptionOptions holds configuration for keystroke captioning.
+type CaptionOptions struct {
+	Font              string
+	FontSize          int
+	KeyStyle          KeyStyle
+	MaxKeysOnscreen   int
+	InactivityTimerMs int
+	Alignment         CaptionAlignment
+	FontColor         string // #RRGGBB hex
+	HighlightColor    string // #RRGGBB hex
+	BoxColor          string // #RRGGBB hex
+	BoxOpacity        float64
+	BoxPadding        int
+	MarginLeft        int
+	MarginRight       int
+	MarginVertical    int
+}
+
+// DefaultCaptionOptions returns caption options with sensible defaults.
+func DefaultCaptionOptions() CaptionOptions {
+	font := defaultOSFont()
+	return CaptionOptions{
+		Font:              font,
+		FontSize:          22,
+		KeyStyle:          KeyStyleIcon,
+		MaxKeysOnscreen:   10,
+		InactivityTimerMs: 1000,
+		Alignment:         AlignBottomCenter,
+		FontColor:         "#FFFFFF",
+		HighlightColor:    "#FFCC66",
+		BoxColor:          "#000000",
+		BoxOpacity:        0.5,
+		BoxPadding:        10,
+		MarginLeft:        20,
+		MarginRight:       20,
+		MarginVertical:    20,
+	}
+}
+
+func defaultOSFont() string {
+	font := "monospace"
+	switch runtime.GOOS {
+	case "windows":
+		font = "Consolas"
+	case "darwin":
+		font = "Menlo"
+	}
+	return font
+}
+
+// KeyStyle represents the rendering style for key captions.
+type KeyStyle string
+
+const (
+	KeyStyleVim  KeyStyle = "vim"
+	KeyStyleIcon KeyStyle = "icon"
+)
+
+// CaptionAlignment controls where captions appear on screen.
+type CaptionAlignment string
+
+const (
+	AlignBottomLeft   CaptionAlignment = "bottom-left"
+	AlignBottomCenter CaptionAlignment = "bottom-center"
+	AlignBottomRight  CaptionAlignment = "bottom-right"
+	AlignMiddleLeft   CaptionAlignment = "middle-left"
+	AlignMiddleCenter CaptionAlignment = "middle-center"
+	AlignMiddleRight  CaptionAlignment = "middle-right"
+	AlignTopLeft      CaptionAlignment = "top-left"
+	AlignTopCenter    CaptionAlignment = "top-center"
+	AlignTopRight     CaptionAlignment = "top-right"
+)
+
+// CaptionAlignments maps alignment names to ASS numpad values (1–9).
+var CaptionAlignments = map[CaptionAlignment]int{
+	AlignBottomLeft:   1,
+	AlignBottomCenter: 2,
+	AlignBottomRight:  3,
+	AlignMiddleLeft:   4,
+	AlignMiddleCenter: 5,
+	AlignMiddleRight:  6,
+	AlignTopLeft:      7,
+	AlignTopCenter:    8,
+	AlignTopRight:     9,
+}
+
+// ASSValue returns the ASS numpad alignment integer (1–9).
+func (a CaptionAlignment) ASSValue() int {
+	if v, ok := CaptionAlignments[a]; ok {
+		return v
+	}
+	return CaptionAlignments[AlignBottomRight]
+}
+
+// Normalize transforms a key name (e.g. "Ctrl+Shift+a") into styled display text.
+func (s KeyStyle) Normalize(key string) string {
+	parts := strings.Split(key, "+")
+	overrides := vimOverrides
+	if s == KeyStyleIcon {
+		overrides = iconOverrides
+	}
+	for i, p := range parts {
+		if o, ok := overrides[p]; ok {
+			parts[i] = o
+		}
+	}
+	if s == KeyStyleIcon {
+		return strings.Join(parts, "")
+	}
+	if len(parts) == 1 && len([]rune(parts[0])) == 1 {
+		return parts[0]
+	}
+	return "<" + strings.Join(parts, "-") + ">"
+}
+
+var vimOverrides = map[string]string{
+	"Backspace": "BS",
+	"Delete":    "Del",
+	"Enter":     "CR",
+	"Escape":    "Esc",
+	"Ctrl":      "C",
+	"Alt":       "A",
+	"Shift":     "S",
+}
+
+var iconOverrides = map[string]string{
+	"Backspace": "⌫",
+	"Delete":    "⌦",
+	"Ctrl":      "^",
+	"Alt":       "⌥",
+	"Shift":     "⇧",
+	"Down":      "↓",
+	"PageDown":  "PgDn",
+	"Up":        "↑",
+	"PageUp":    "PgUp",
+	"Left":      "←",
+	"Right":     "→",
+	"Space":     "␣",
+	"Enter":     "↵",
+	"Escape":    "Esc",
+	"Tab":       "⇥",
+}
+
+// CaptionWindow describes the sliding key display at the moment each key is pressed.
+type CaptionWindow struct {
+	First, Last int
+	ShowUntil   int
+	IsTruncated bool
+}
+
+// captionWindows returns one CaptionWindow per event. Windows are computed
+// such that keys remain on screen until inactivityTimerMs has been exceeded.
+// Further, there will be no more than maxKeys shown on the screen at any
+// point in time.
+func captionWindows(events []KeyEvent, inactivityTimerMs, maxKeys int) []CaptionWindow {
+	n := len(events)
+	wins := make([]CaptionWindow, 0, n)
+	sessionStart := 0
+
+	for i := range n {
+		current := events[i]
+		windowStart := max(sessionStart, i-maxKeys+1)
+		isTruncated := windowStart > sessionStart
+
+		var showUntil int
+		if i+1 < n && int(events[i+1].StartMs)-int(current.StartMs) < inactivityTimerMs {
+			showUntil = int(events[i+1].StartMs)
+		} else {
+			showUntil = int(current.StartMs) + inactivityTimerMs
+			sessionStart = i + 1
+		}
+
+		wins = append(wins, CaptionWindow{
+			First:       windowStart,
+			Last:        i,
+			ShowUntil:   showUntil,
+			IsTruncated: isTruncated,
+		})
+	}
+	return wins
+}
+
+// ASS helpers
+
+func msToASS(ms int) string {
+	h := ms / 3600000
+	ms %= 3600000
+	m := ms / 60000
+	ms %= 60000
+	s := ms / 1000
+	ms %= 1000
+	cs := ms / 10
+	return fmt.Sprintf("%d:%02d:%02d.%02d", h, m, s, cs)
+}
+
+func assEscape(s string) string {
+	s = strings.ReplaceAll(s, `\`, `\\`)
+	// Restore ASS newline sequences that were double-escaped
+	s = strings.ReplaceAll(s, `\\N`, `\N`)
+	s = strings.ReplaceAll(s, `\\n`, `\n`)
+	s = strings.ReplaceAll(s, "{", `\{`)
+	s = strings.ReplaceAll(s, "}", `\}`)
+	return s
+}
+
+func opacityToASSAlpha(opacity float64) string {
+	alpha := int(math.Round((1 - opacity) * 255))
+	return fmt.Sprintf("%02X", alpha)
+}
+
+func hexToASSBGR(hex string) string {
+	hex = strings.TrimPrefix(hex, "#")
+	r, g, b := hex[0:2], hex[2:4], hex[4:6]
+	return fmt.Sprintf("&H%s%s%s&", b, g, r)
+}
+
+// captionWindowPlainText builds plain text with no inline overrides.
+// Used for the background box layer so the opaque box renders without seams.
+func captionWindowPlainText(events []KeyEvent, w CaptionWindow) string {
+	var parts []string
+	if w.IsTruncated {
+		parts = append(parts, "…")
+	}
+	for i := w.First; i <= w.Last; i++ {
+		parts = append(parts, assEscape(events[i].Key))
+	}
+	return strings.Join(parts, " ")
+}
+
+// captionWindowColoredText builds ASS text with the last key highlighted.
+// Used for the visible text layer (no opaque box, so color overrides cause no seams).
+func captionWindowColoredText(events []KeyEvent, w CaptionWindow, highlightColor string) string {
+	var parts []string
+	if w.IsTruncated {
+		parts = append(parts, "…")
+	}
+	for i := w.First; i < w.Last; i++ {
+		parts = append(parts, assEscape(events[i].Key))
+	}
+	lastKey := assEscape(events[w.Last].Key)
+	parts = append(parts, fmt.Sprintf(`{\c%s}%s`, highlightColor, lastKey))
+	return strings.Join(parts, " ")
+}
+
+const ASSHeaderTemplate = `[Script Info]
+ScriptType: v4.00+
+PlayResX: {{.ResX}}
+PlayResY: {{.ResY}}
+ScaledBorderAndShadow: yes
+
+[V4+ Styles]
+Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
+{{- if .HasCaption}}
+Style: KeysBG,{{.Font}},{{.FontSize}},&H00FFFFFF,&H000000FF,{{.BoxColor}},&H00000000,1,0,0,0,100,100,0,0,3,{{.BoxPadding}},0,{{.Alignment}},{{.MarginLeft}},{{.MarginRight}},{{.MarginVertical}},1
+Style: KeysFG,{{.Font}},{{.FontSize}},{{.FontColor}},&H000000FF,&H00000000,&H00000000,1,0,0,0,100,100,0,0,1,0,0,{{.Alignment}},{{.MarginLeft}},{{.MarginRight}},{{.MarginVertical}},1
+{{- end}}
+{{- if .HasOverlay}}
+Style: OverlayBG,{{.OverlayFont}},{{.OverlayFontSize}},&H00FFFFFF,&H000000FF,{{.OverlayBoxColor}},&H00000000,1,0,0,0,100,100,0,0,3,{{.OverlayBoxPadding}},0,{{.OverlayAlignment}},{{.OverlayMarginLeft}},{{.OverlayMarginRight}},{{.OverlayMarginVertical}},1
+Style: OverlayFG,{{.OverlayFont}},{{.OverlayFontSize}},{{.OverlayFontColor}},&H000000FF,&H00000000,&H00000000,1,0,0,0,100,100,0,0,1,0,0,{{.OverlayAlignment}},{{.OverlayMarginLeft}},{{.OverlayMarginRight}},{{.OverlayMarginVertical}},1
+{{- end}}
+
+[Events]
+Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text`
+
+type ASSHeaderData struct {
+	ResX, ResY            int
+	Font                  string
+	FontColor             string
+	FontSize              int
+	Alignment             int
+	MarginLeft            int
+	MarginRight           int
+	MarginVertical        int
+	BoxColor              string
+	BoxPadding            int
+	HasCaption            bool
+	HasOverlay            bool
+	OverlayFont           string
+	OverlayFontColor      string
+	OverlayFontSize       int
+	OverlayAlignment      int
+	OverlayMarginLeft     int
+	OverlayMarginRight    int
+	OverlayMarginVertical int
+	OverlayBoxColor       string
+	OverlayBoxPadding     int
+}
+
+var ASSHeader = template.Must(template.New("captionASS").Parse(ASSHeaderTemplate))
+
+// GenerateCaptionFile creates an ASS subtitle file from key events and/or overlay events and returns its path.
+func GenerateCaptionFile(events []KeyEvent, overlays []OverlayEvent, videoOpts VideoOptions, opts CaptionOptions, overlayOpts OverlayOptions) (string, error) {
+	if len(events) == 0 && len(overlays) == 0 {
+		return "", nil
+	}
+
+	width, height := calcTermDimensions(*videoOpts.Style)
+	playbackSpeed := videoOpts.PlaybackSpeed
+	tempDir := videoOpts.Input
+
+	// Normalize caption events: apply key style and adjust for playback speed.
+	normalized := make([]KeyEvent, len(events))
+	for i, e := range events {
+		normalized[i] = KeyEvent{
+			StartMs: e.StartMs,
+			Key:     opts.KeyStyle.Normalize(e.Key),
+		}
+	}
+
+	// Normalize overlay events: copy so we can adjust timestamps in place.
+	normalizedOverlays := make([]OverlayEvent, len(overlays))
+	copy(normalizedOverlays, overlays)
+
+	// Adjust all timestamps for playback speed.
+	if playbackSpeed != 0 && playbackSpeed != 1.0 {
+		for i := range normalized {
+			normalized[i].StartMs = int64(float64(normalized[i].StartMs) / playbackSpeed)
+		}
+		for i := range normalizedOverlays {
+			normalizedOverlays[i].StartMs = int64(float64(normalizedOverlays[i].StartMs) / playbackSpeed)
+			normalizedOverlays[i].DurationMs = int64(float64(normalizedOverlays[i].DurationMs) / playbackSpeed)
+		}
+	}
+
+	assPath := filepath.Join(tempDir, "captions.ass")
+	f, err := os.Create(assPath)
+	if err != nil {
+		return "", fmt.Errorf("failed to create caption file: %w", err)
+	}
+	defer f.Close() //nolint:errcheck
+
+	data := ASSHeaderData{
+		ResX:                  width,
+		ResY:                  height,
+		HasCaption:            len(events) > 0,
+		HasOverlay:            len(overlays) > 0,
+		Font:                  opts.Font,
+		FontSize:              opts.FontSize,
+		Alignment:             opts.Alignment.ASSValue(),
+		MarginLeft:            opts.MarginLeft,
+		MarginRight:           opts.MarginRight,
+		MarginVertical:        opts.MarginVertical,
+		BoxPadding:            opts.BoxPadding,
+		FontColor:             hexToASSBGR(opts.FontColor),
+		BoxColor:              hexToASSBGR(opts.BoxColor),
+		OverlayFont:           overlayOpts.Font,
+		OverlayFontSize:       overlayOpts.FontSize,
+		OverlayAlignment:      overlayOpts.Alignment.ASSValue(),
+		OverlayMarginLeft:     overlayOpts.MarginLeft,
+		OverlayMarginRight:    overlayOpts.MarginRight,
+		OverlayMarginVertical: overlayOpts.MarginVertical,
+		OverlayBoxPadding:     overlayOpts.BoxPadding,
+		OverlayFontColor:      hexToASSBGR(overlayOpts.FontColor),
+		OverlayBoxColor:       hexToASSBGR(overlayOpts.BoxColor),
+	}
+
+	if err := ASSHeader.Execute(f, data); err != nil {
+		return "", fmt.Errorf("failed to write caption header: %w", err)
+	}
+	_, _ = fmt.Fprintln(f)
+
+	alpha := opacityToASSAlpha(opts.BoxOpacity)
+	highlightColor := hexToASSBGR(opts.HighlightColor)
+
+	for _, w := range captionWindows(normalized, opts.InactivityTimerMs, max(opts.MaxKeysOnscreen, 1)) {
+		start := msToASS(int(normalized[w.Last].StartMs))
+		end := msToASS(w.ShowUntil)
+
+		// Layer 0: invisible text with seamless opaque box (no inline overrides = no seams)
+		plainText := captionWindowPlainText(normalized, w)
+		bgLine := fmt.Sprintf(`Dialogue: 0,%s,%s,KeysBG,,0,0,0,,{\3a&H%s&\1a&HFF&}%s`,
+			start, end, alpha, plainText)
+		_, _ = fmt.Fprintln(f, bgLine)
+
+		// Layer 1: visible colored text with no box
+		coloredText := captionWindowColoredText(normalized, w, highlightColor)
+		fgLine := fmt.Sprintf(`Dialogue: 1,%s,%s,KeysFG,,0,0,0,,%s`,
+			start, end, coloredText)
+		_, _ = fmt.Fprintln(f, fgLine)
+	}
+
+	overlayAlpha := opacityToASSAlpha(overlayOpts.BoxOpacity)
+
+	for _, o := range normalizedOverlays {
+		start := msToASS(int(o.StartMs))
+		end := msToASS(int(o.StartMs + o.DurationMs))
+		text := assEscape(o.Text)
+
+		// Layer 2: invisible text with opaque box
+		bgLine := fmt.Sprintf(`Dialogue: 2,%s,%s,OverlayBG,,0,0,0,,{\3a&H%s&\1a&HFF&}%s`,
+			start, end, overlayAlpha, text)
+		_, _ = fmt.Fprintln(f, bgLine)
+
+		// Layer 3: visible text
+		fgLine := fmt.Sprintf(`Dialogue: 3,%s,%s,OverlayFG,,0,0,0,,%s`,
+			start, end, text)
+		_, _ = fmt.Fprintln(f, fgLine)
+	}
+
+	return assPath, nil
+}
diff --git a/caption_test.go b/caption_test.go
new file mode 100644
index 00000000..4b0912d8
--- /dev/null
+++ b/caption_test.go
@@ -0,0 +1,444 @@
+package main
+
+import (
+	"fmt"
+	"os"
+	"path/filepath"
+	"reflect"
+	"strings"
+	"testing"
+)
+
+func keyEvents(ms ...int64) []KeyEvent {
+	out := make([]KeyEvent, len(ms))
+	for i, t := range ms {
+		out[i] = KeyEvent{StartMs: t, Key: "a"}
+	}
+	return out
+}
+
+func TestVimNormalize(t *testing.T) {
+	tests := []struct {
+		input string
+		want  string
+	}{
+		{"a", "a"},
+		{"Enter", "<CR>"},
+		{"Backspace", "<BS>"},
+		{"Escape", "<Esc>"},
+		{"Ctrl+c", "<C-c>"},
+		{"Alt+f", "<A-f>"},
+		{"Shift+Tab", "<S-Tab>"},
+		{"Space", "<Space>"},
+		{"Delete", "<Del>"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.input, func(t *testing.T) {
+			got := KeyStyleVim.Normalize(tt.input)
+			if got != tt.want {
+				t.Errorf("KeyStyleVim.Normalize(%q) = %q, want %q", tt.input, got, tt.want)
+			}
+		})
+	}
+}
+
+func TestIconNormalize(t *testing.T) {
+	tests := []struct {
+		input string
+		want  string
+	}{
+		{"a", "a"},
+		{"Enter", "↵"},
+		{"Escape", "Esc"},
+		{"PageUp", "PgUp"},
+		{"PageDown", "PgDn"},
+		{"Backspace", "⌫"},
+		{"Space", "␣"},
+		{"Ctrl+c", "^c"},
+		{"Alt+d", "⌥d"},
+		{"Shift+Tab", "⇧⇥"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.input, func(t *testing.T) {
+			got := KeyStyleIcon.Normalize(tt.input)
+			if got != tt.want {
+				t.Errorf("KeyStyleIcon.Normalize(%q) = %q, want %q", tt.input, got, tt.want)
+			}
+		})
+	}
+}
+
+func TestCaptionWindows(t *testing.T) {
+	tests := []struct {
+		name              string
+		events            []KeyEvent
+		inactivityTimerMs int
+		maxKeys           int
+		want              []CaptionWindow
+	}{
+		{
+			name:              "empty events",
+			events:            []KeyEvent{},
+			inactivityTimerMs: 1000,
+			maxKeys:           3,
+			want:              []CaptionWindow{},
+		},
+		{
+			name:              "single event",
+			events:            keyEvents(0),
+			inactivityTimerMs: 1000,
+			maxKeys:           3,
+			want: []CaptionWindow{
+				{First: 0, Last: 0, ShowUntil: 1000, IsTruncated: false},
+			},
+		},
+		{
+			name:              "two events in same session",
+			events:            keyEvents(0, 500),
+			inactivityTimerMs: 1000,
+			maxKeys:           3,
+			want: []CaptionWindow{
+				{First: 0, Last: 0, ShowUntil: 500, IsTruncated: false},
+				{First: 0, Last: 1, ShowUntil: 1500, IsTruncated: false},
+			},
+		},
+		{
+			name:              "gap exactly equals inactivity timer starts new session",
+			events:            keyEvents(0, 1000),
+			inactivityTimerMs: 1000,
+			maxKeys:           3,
+			want: []CaptionWindow{
+				{First: 0, Last: 0, ShowUntil: 1000, IsTruncated: false},
+				{First: 1, Last: 1, ShowUntil: 2000, IsTruncated: false},
+			},
+		},
+		{
+			name:              "gap one over inactivity timer starts new session",
+			events:            keyEvents(0, 1001),
+			inactivityTimerMs: 1000,
+			maxKeys:           3,
+			want: []CaptionWindow{
+				{First: 0, Last: 0, ShowUntil: 1000, IsTruncated: false},
+				{First: 1, Last: 1, ShowUntil: 2001, IsTruncated: false},
+			},
+		},
+		{
+			name:              "sliding window truncates when maxKeys exceeded",
+			events:            keyEvents(0, 100, 200, 300),
+			inactivityTimerMs: 1000,
+			maxKeys:           3,
+			want: []CaptionWindow{
+				{First: 0, Last: 0, ShowUntil: 100, IsTruncated: false},
+				{First: 0, Last: 1, ShowUntil: 200, IsTruncated: false},
+				{First: 0, Last: 2, ShowUntil: 300, IsTruncated: false},
+				{First: 1, Last: 3, ShowUntil: 1300, IsTruncated: true},
+			},
+		},
+		{
+			name:              "session reset clears truncation",
+			events:            keyEvents(0, 100, 200, 2000, 2100),
+			inactivityTimerMs: 1000,
+			maxKeys:           2,
+			want: []CaptionWindow{
+				{First: 0, Last: 0, ShowUntil: 100, IsTruncated: false},
+				{First: 0, Last: 1, ShowUntil: 200, IsTruncated: false},
+				{First: 1, Last: 2, ShowUntil: 1200, IsTruncated: true},
+				{First: 3, Last: 3, ShowUntil: 2100, IsTruncated: false},
+				{First: 3, Last: 4, ShowUntil: 3100, IsTruncated: false},
+			},
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := captionWindows(tt.events, tt.inactivityTimerMs, tt.maxKeys)
+			if len(got) == 0 && len(tt.want) == 0 {
+				return
+			}
+			if !reflect.DeepEqual(got, tt.want) {
+				t.Errorf("got  %v\nwant %v", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestMsToASS(t *testing.T) {
+	tests := []struct {
+		ms   int
+		want string
+	}{
+		{0, "0:00:00.00"},
+		{1000, "0:00:01.00"},
+		{61000, "0:01:01.00"},
+		{3661050, "1:01:01.05"},
+		{500, "0:00:00.50"},
+		{10, "0:00:00.01"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.want, func(t *testing.T) {
+			got := msToASS(tt.ms)
+			if got != tt.want {
+				t.Errorf("msToASS(%d) = %q, want %q", tt.ms, got, tt.want)
+			}
+		})
+	}
+}
+
+func TestAssEscape(t *testing.T) {
+	tests := []struct {
+		input string
+		want  string
+	}{
+		{`hello`, `hello`},
+		{`a\b`, `a\\b`},
+		{`{tag}`, `\{tag\}`},
+		{`a\{b}`, `a\\\{b\}`},
+		{`a\Nb`, `a\Nb`},
+	}
+	for _, tt := range tests {
+		t.Run(tt.input, func(t *testing.T) {
+			got := assEscape(tt.input)
+			if got != tt.want {
+				t.Errorf("assEscape(%q) = %q, want %q", tt.input, got, tt.want)
+			}
+		})
+	}
+}
+
+func TestOpacityToASSAlpha(t *testing.T) {
+	tests := []struct {
+		opacity float64
+		want    string
+	}{
+		{1.0, "00"},
+		{0.0, "FF"},
+		{0.5, "80"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.want, func(t *testing.T) {
+			got := opacityToASSAlpha(tt.opacity)
+			if got != tt.want {
+				t.Errorf("opacityToASSAlpha(%f) = %q, want %q", tt.opacity, got, tt.want)
+			}
+		})
+	}
+}
+
+func TestHexToASSBGR(t *testing.T) {
+	tests := []struct {
+		hex  string
+		want string
+	}{
+		{"#FFCC66", "&H66CCFF&"},
+		{"#FF0000", "&H0000FF&"},
+		{"#00FF00", "&H00FF00&"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.hex, func(t *testing.T) {
+			got := hexToASSBGR(tt.hex)
+			if got != tt.want {
+				t.Errorf("hexToASSBGR(%q) = %q, want %q", tt.hex, got, tt.want)
+			}
+		})
+	}
+}
+
+func TestCaptionAlignment(t *testing.T) {
+	tests := []struct {
+		alignment CaptionAlignment
+		want      int
+	}{
+		{AlignBottomLeft, 1},
+		{AlignBottomCenter, 2},
+		{AlignBottomRight, 3},
+		{AlignMiddleLeft, 4},
+		{AlignMiddleCenter, 5},
+		{AlignMiddleRight, 6},
+		{AlignTopLeft, 7},
+		{AlignTopCenter, 8},
+		{AlignTopRight, 9},
+		{"unknown", 3}, // default
+	}
+	for _, tt := range tests {
+		t.Run(string(tt.alignment), func(t *testing.T) {
+			got := tt.alignment.ASSValue()
+			if got != tt.want {
+				t.Errorf("CaptionAlignment(%q).ASSValue() = %d, want %d", tt.alignment, got, tt.want)
+			}
+		})
+	}
+}
+
+func TestGenerateCaptionFile(t *testing.T) {
+	tmpDir := t.TempDir()
+
+	events := []KeyEvent{
+		{StartMs: 0, Key: "h"},
+		{StartMs: 100, Key: "i"},
+		{StartMs: 200, Key: "Enter"},
+	}
+
+	videoOpts := VideoOptions{PlaybackSpeed: 1.0, Input: tmpDir, Style: &StyleOptions{Width: 800, Height: 600}}
+	path, err := GenerateCaptionFile(events, nil, videoOpts, DefaultCaptionOptions(), DefaultOverlayOptions())
+	if err != nil {
+		t.Fatalf("GenerateCaptionFile failed: %v", err)
+	}
+
+	if filepath.Dir(path) != tmpDir {
+		t.Errorf("expected file in %s, got %s", tmpDir, path)
+	}
+
+	data, err := os.ReadFile(path)
+	if err != nil {
+		t.Fatalf("failed to read caption file: %v", err)
+	}
+	content := string(data)
+
+	if !strings.Contains(content, "[Script Info]") {
+		t.Error("expected ASS header in output")
+	}
+	if !strings.Contains(content, "PlayResX: 800") {
+		t.Error("expected PlayResX: 800")
+	}
+	if !strings.Contains(content, "Dialogue:") {
+		t.Error("expected Dialogue lines in output")
+	}
+	expectedFont := DefaultCaptionOptions().Font
+	if !strings.Contains(content, expectedFont) {
+		t.Errorf("expected default font %q in output", expectedFont)
+	}
+	// Default alignment is bottom-center (ASS numpad 2)
+	if !strings.Contains(content, ",2,20,20,20,") {
+		t.Error("expected default alignment value 2 in style lines")
+	}
+}
+
+func TestGenerateCaptionFileDefaultFont(t *testing.T) {
+	tmpDir := t.TempDir()
+
+	events := []KeyEvent{{StartMs: 0, Key: "a"}}
+
+	// When CaptionFont is empty, should default to the OS-appropriate font
+	videoOpts := VideoOptions{PlaybackSpeed: 1.0, Input: tmpDir, Style: &StyleOptions{Width: 800, Height: 600}}
+	path, err := GenerateCaptionFile(events, nil, videoOpts, DefaultCaptionOptions(), DefaultOverlayOptions())
+	if err != nil {
+		t.Fatalf("GenerateCaptionFile failed: %v", err)
+	}
+
+	data, err := os.ReadFile(path)
+	if err != nil {
+		t.Fatalf("failed to read caption file: %v", err)
+	}
+	content := string(data)
+
+	expectedFont := DefaultCaptionOptions().Font
+	if !strings.Contains(content, expectedFont+",") {
+		t.Errorf("expected default font %q in ASS output", expectedFont)
+	}
+}
+
+func TestGenerateCaptionFileExplicitFontOverride(t *testing.T) {
+	tmpDir := t.TempDir()
+
+	events := []KeyEvent{{StartMs: 0, Key: "a"}}
+	opts := DefaultCaptionOptions()
+	opts.Font = "JetBrainsMonoNL NFM"
+
+	videoOpts := VideoOptions{PlaybackSpeed: 1.0, Input: tmpDir, Style: &StyleOptions{Width: 800, Height: 600}}
+	path, err := GenerateCaptionFile(events, nil, videoOpts, opts, DefaultOverlayOptions())
+	if err != nil {
+		t.Fatalf("GenerateCaptionFile failed: %v", err)
+	}
+
+	data, err := os.ReadFile(path)
+	if err != nil {
+		t.Fatalf("failed to read caption file: %v", err)
+	}
+	content := string(data)
+
+	if !strings.Contains(content, "JetBrainsMonoNL NFM") {
+		t.Error("expected explicit CaptionFont in ASS output")
+	}
+	defaultFont := DefaultCaptionOptions().Font
+	if strings.Contains(content, defaultFont) {
+		t.Errorf("default font %q should not appear when CaptionFont is set explicitly", defaultFont)
+	}
+}
+
+func TestGenerateCaptionFilePlaybackSpeed(t *testing.T) {
+	tmpDir := t.TempDir()
+
+	events := []KeyEvent{
+		{StartMs: 0, Key: "a"},
+		{StartMs: 2000, Key: "b"},
+	}
+
+	videoOpts := VideoOptions{PlaybackSpeed: 2.0, Input: tmpDir, Style: &StyleOptions{Width: 800, Height: 600}}
+	path, err := GenerateCaptionFile(events, nil, videoOpts, DefaultCaptionOptions(), DefaultOverlayOptions())
+	if err != nil {
+		t.Fatalf("GenerateCaptionFile failed: %v", err)
+	}
+
+	data, err := os.ReadFile(path)
+	if err != nil {
+		t.Fatalf("failed to read caption file: %v", err)
+	}
+	content := string(data)
+
+	// At 2x speed, 2000ms becomes 1000ms = 0:00:01.00
+	if !strings.Contains(content, "0:00:01.00") {
+		t.Error("expected timestamp adjusted for 2x playback speed")
+	}
+}
+
+func TestCaptionFontSizeConsistentAcrossHeights(t *testing.T) {
+	fontSize := 22
+	var firstContent string
+	for _, height := range []int{384, 600, 768, 1080} {
+		t.Run(fmt.Sprintf("height_%d", height), func(t *testing.T) {
+			tmpDir := t.TempDir()
+			events := []KeyEvent{{StartMs: 0, Key: "a"}}
+			opts := DefaultCaptionOptions()
+			opts.FontSize = fontSize
+
+			videoOpts := VideoOptions{
+				PlaybackSpeed: 1.0,
+				Input:         tmpDir,
+				Style:         &StyleOptions{Width: 800, Height: height},
+			}
+			path, err := GenerateCaptionFile(events, nil, videoOpts, opts, DefaultOverlayOptions())
+			if err != nil {
+				t.Fatalf("GenerateCaptionFile failed: %v", err)
+			}
+			data, err := os.ReadFile(path)
+			if err != nil {
+				t.Fatalf("failed to read caption file: %v", err)
+			}
+			content := string(data)
+			// Extract the font size from the style line
+			for _, line := range strings.Split(content, "\n") {
+				if strings.HasPrefix(line, "Style: KeysFG,") {
+					if firstContent == "" {
+						firstContent = line
+					} else if line != firstContent {
+						t.Errorf("font size differs across heights:\n  first:   %s\n  current: %s", firstContent, line)
+					}
+					break
+				}
+			}
+		})
+	}
+}
+
+func TestGenerateCaptionFileEmpty(t *testing.T) {
+	tmpDir := t.TempDir()
+
+	videoOpts := VideoOptions{PlaybackSpeed: 1.0, Input: tmpDir, Style: &StyleOptions{Width: 800, Height: 600}}
+	path, err := GenerateCaptionFile([]KeyEvent{}, nil, videoOpts, DefaultCaptionOptions(), DefaultOverlayOptions())
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if path != "" {
+		t.Errorf("expected empty path for no events, got %q", path)
+	}
+}
diff --git a/command.go b/command.go
index 4e5b0d06..ae059311 100644
--- a/command.go
+++ b/command.go
@@ -8,6 +8,7 @@ import (
 	"regexp"
 	"strconv"
 	"strings"
+	"sync/atomic"
 	"time"
 
 	"github.com/atotto/clipboard"
@@ -70,6 +71,9 @@ var CommandFuncs = map[parser.CommandType]CommandFunc{
 	token.PASTE:       ExecutePaste,
 	token.ENV:         ExecuteEnv,
 	token.WAIT:        ExecuteWait,
+	token.OVERLAY:     ExecuteOverlay,
+	token.CAPTION_ON:  ExecuteCaptionOn,
+	token.CAPTION_OFF: ExecuteCaptionOff,
 }
 
 // ExecuteNoop is a no-op command that does nothing.
@@ -94,6 +98,7 @@ func ExecuteKey(k input.Key) CommandFunc {
 			repeat = 1
 		}
 		for i := 0; i < repeat; i++ {
+			v.KeyLogger.LogKey(c.Type.String())
 			err = v.Page.Keyboard.Type(k)
 			if err != nil {
 				return fmt.Errorf("failed to type key %c: %w", k, err)
@@ -211,6 +216,7 @@ func ExecuteWait(c parser.Command, v *VHS) error {
 // ExecuteCtrl is a CommandFunc that presses the argument keys and/or modifiers
 // with the ctrl key held down on the running instance of vhs.
 func ExecuteCtrl(c parser.Command, v *VHS) error {
+	v.KeyLogger.LogKey("Ctrl+" + c.Args)
 	// Create key combination by holding ControlLeft
 	action := v.Page.KeyActions().Press(input.ControlLeft)
 	keys := strings.Split(c.Args, " ")
@@ -266,6 +272,7 @@ func ExecuteCtrl(c parser.Command, v *VHS) error {
 // ExecuteAlt is a CommandFunc that presses the argument key with the alt key
 // held down on the running instance of vhs.
 func ExecuteAlt(c parser.Command, v *VHS) error {
+	v.KeyLogger.LogKey("Alt+" + c.Args)
 	err := v.Page.Keyboard.Press(input.AltLeft)
 	if err != nil {
 		return fmt.Errorf("failed to press Alt key: %w", err)
@@ -305,6 +312,7 @@ func ExecuteAlt(c parser.Command, v *VHS) error {
 // ExecuteShift is a CommandFunc that presses the argument key with the shift
 // key held down on the running instance of vhs.
 func ExecuteShift(c parser.Command, v *VHS) error {
+	v.KeyLogger.LogKey("Shift+" + c.Args)
 	err := v.Page.Keyboard.Press(input.ShiftLeft)
 	if err != nil {
 		return fmt.Errorf("failed to press Shift key: %w", err)
@@ -383,6 +391,11 @@ func ExecuteType(c parser.Command, v *VHS) error {
 		}
 	}
 	for _, r := range c.Args {
+		if r == ' ' { // make in consistent in the log file
+			v.KeyLogger.LogKey("Space")
+		} else {
+			v.KeyLogger.LogKey(string(r))
+		}
 		k, ok := keymap[r]
 		if ok {
 			err := v.Page.Keyboard.Type(k)
@@ -458,27 +471,51 @@ func ExecutePaste(_ parser.Command, v *VHS) error {
 
 // Settings maps the Set commands to their respective functions.
 var Settings = map[string]CommandFunc{
-	"FontFamily":    ExecuteSetFontFamily,
-	"FontSize":      ExecuteSetFontSize,
-	"Framerate":     ExecuteSetFramerate,
-	"Height":        ExecuteSetHeight,
-	"LetterSpacing": ExecuteSetLetterSpacing,
-	"LineHeight":    ExecuteSetLineHeight,
-	"PlaybackSpeed": ExecuteSetPlaybackSpeed,
-	"Padding":       ExecuteSetPadding,
-	"Theme":         ExecuteSetTheme,
-	"TypingSpeed":   ExecuteSetTypingSpeed,
-	"Width":         ExecuteSetWidth,
-	"Shell":         ExecuteSetShell,
-	"LoopOffset":    ExecuteLoopOffset,
-	"MarginFill":    ExecuteSetMarginFill,
-	"Margin":        ExecuteSetMargin,
-	"WindowBar":     ExecuteSetWindowBar,
-	"WindowBarSize": ExecuteSetWindowBarSize,
-	"BorderRadius":  ExecuteSetBorderRadius,
-	"WaitPattern":   ExecuteSetWaitPattern,
-	"WaitTimeout":   ExecuteSetWaitTimeout,
-	"CursorBlink":   ExecuteSetCursorBlink,
+	"FontFamily":             ExecuteSetFontFamily,
+	"FontSize":               ExecuteSetFontSize,
+	"Framerate":              ExecuteSetFramerate,
+	"Height":                 ExecuteSetHeight,
+	"LetterSpacing":          ExecuteSetLetterSpacing,
+	"LineHeight":             ExecuteSetLineHeight,
+	"PlaybackSpeed":          ExecuteSetPlaybackSpeed,
+	"Padding":                ExecuteSetPadding,
+	"Theme":                  ExecuteSetTheme,
+	"TypingSpeed":            ExecuteSetTypingSpeed,
+	"Width":                  ExecuteSetWidth,
+	"Shell":                  ExecuteSetShell,
+	"LoopOffset":             ExecuteLoopOffset,
+	"MarginFill":             ExecuteSetMarginFill,
+	"Margin":                 ExecuteSetMargin,
+	"WindowBar":              ExecuteSetWindowBar,
+	"WindowBarSize":          ExecuteSetWindowBarSize,
+	"BorderRadius":           ExecuteSetBorderRadius,
+	"WaitPattern":            ExecuteSetWaitPattern,
+	"WaitTimeout":            ExecuteSetWaitTimeout,
+	"CursorBlink":            ExecuteSetCursorBlink,
+	"CaptionFont":            ExecuteSetCaptionFont,
+	"CaptionFontSize":        ExecuteSetCaptionFontSize,
+	"CaptionKeyStyle":        ExecuteSetCaptionKeyStyle,
+	"CaptionMaxKeys":         ExecuteSetCaptionMaxKeys,
+	"CaptionInactivityTimer": ExecuteSetCaptionInactivityTimer,
+	"CaptionAlignment":       ExecuteSetCaptionAlignment,
+	"CaptionFontColor":       ExecuteSetCaptionFontColor,
+	"CaptionHighlightColor":  ExecuteSetCaptionHighlightColor,
+	"CaptionBoxColor":        ExecuteSetCaptionBoxColor,
+	"CaptionBoxOpacity":      ExecuteSetCaptionBoxOpacity,
+	"CaptionBoxPadding":      ExecuteSetCaptionBoxPadding,
+	"CaptionMarginLeft":      ExecuteSetCaptionMarginLeft,
+	"CaptionMarginRight":     ExecuteSetCaptionMarginRight,
+	"CaptionMarginVertical":  ExecuteSetCaptionMarginVertical,
+	"OverlayFont":            ExecuteSetOverlayFont,
+	"OverlayFontSize":        ExecuteSetOverlayFontSize,
+	"OverlayFontColor":       ExecuteSetOverlayFontColor,
+	"OverlayBoxColor":        ExecuteSetOverlayBoxColor,
+	"OverlayBoxOpacity":      ExecuteSetOverlayBoxOpacity,
+	"OverlayBoxPadding":      ExecuteSetOverlayBoxPadding,
+	"OverlayAlignment":       ExecuteSetOverlayAlignment,
+	"OverlayMarginLeft":      ExecuteSetOverlayMarginLeft,
+	"OverlayMarginRight":     ExecuteSetOverlayMarginRight,
+	"OverlayMarginVertical":  ExecuteSetOverlayMarginVertical,
 }
 
 // ExecuteSet applies the settings on the running vhs specified by the
@@ -754,6 +791,264 @@ func ExecuteScreenshot(c parser.Command, v *VHS) error {
 	return nil
 }
 
+// ensureLibass checks that ffmpeg was compiled with libass support, which is
+// required for CaptionOn. We run `ffmpeg -filters` and look for the
+// "ass" filter rather than letting the render fail late with a cryptic error.
+func ensureLibass() error {
+	// ffmpeg -filters writes to stderr, so capture combined output
+	cmd := exec.Command("ffmpeg", "-filters")
+	out, err := cmd.CombinedOutput()
+	if err != nil {
+		return fmt.Errorf("could not query ffmpeg filters: %w", err)
+	}
+	// Each filter line looks like: " T. ass  V->V  Render ASS subtitles"
+	// Match " ass " with surrounding spaces to avoid false positives (bass, allpass, etc.)
+	for _, line := range strings.Split(string(out), "\n") {
+		fields := strings.Fields(line)
+		if len(fields) >= 2 && fields[1] == "ass" {
+			return nil
+		}
+	}
+	return fmt.Errorf("ffmpeg was not compiled with libass support, which is required for CaptionOn.\nReinstall ffmpeg with libass (e.g. `brew install ffmpeg-full` on macOS)")
+}
+
+// ExecuteCaptionOn enables caption key logging.
+func ExecuteCaptionOn(_ parser.Command, v *VHS) error {
+	if err := ensureLibass(); err != nil {
+		return err
+	}
+	v.KeyLogger.Enable()
+	return nil
+}
+
+// ExecuteCaptionOff disables caption key logging.
+func ExecuteCaptionOff(_ parser.Command, v *VHS) error {
+	v.KeyLogger.Disable()
+	return nil
+}
+
+// ExecuteSetCaptionFont sets the caption font family.
+func ExecuteSetCaptionFont(c parser.Command, v *VHS) error {
+	v.Options.Caption.Font = c.Args
+	return nil
+}
+
+// ExecuteSetCaptionFontSize sets the caption font size.
+func ExecuteSetCaptionFontSize(c parser.Command, v *VHS) error {
+	size, err := strconv.Atoi(c.Args)
+	if err != nil {
+		return fmt.Errorf("failed to parse caption font size: %w", err)
+	}
+	v.Options.Caption.FontSize = size
+	return nil
+}
+
+// ExecuteSetCaptionMaxKeys sets the caption sliding window size.
+func ExecuteSetCaptionMaxKeys(c parser.Command, v *VHS) error {
+	n, err := strconv.Atoi(c.Args)
+	if err != nil {
+		return fmt.Errorf("failed to parse caption max keys: %w", err)
+	}
+	v.Options.Caption.MaxKeysOnscreen = n
+	return nil
+}
+
+// ExecuteSetCaptionInactivityTimer sets the caption inactivity timer.
+func ExecuteSetCaptionInactivityTimer(c parser.Command, v *VHS) error {
+	dur, err := time.ParseDuration(c.Args)
+	if err != nil {
+		return fmt.Errorf("failed to parse caption inactivity timer: %w", err)
+	}
+	v.Options.Caption.InactivityTimerMs = int(dur.Milliseconds())
+	return nil
+}
+
+// ExecuteSetCaptionHighlightColor sets the caption highlight color.
+func ExecuteSetCaptionHighlightColor(c parser.Command, v *VHS) error {
+	v.Options.Caption.HighlightColor = c.Args
+	return nil
+}
+
+// ExecuteSetCaptionFontColor sets the caption text color for non-highlighted keys.
+func ExecuteSetCaptionFontColor(c parser.Command, v *VHS) error {
+	v.Options.Caption.FontColor = c.Args
+	return nil
+}
+
+// ExecuteSetCaptionBoxColor sets the caption background box fill color.
+func ExecuteSetCaptionBoxColor(c parser.Command, v *VHS) error {
+	v.Options.Caption.BoxColor = c.Args
+	return nil
+}
+
+// ExecuteSetCaptionBoxOpacity sets the caption box opacity.
+func ExecuteSetCaptionBoxOpacity(c parser.Command, v *VHS) error {
+	opacity, err := strconv.ParseFloat(c.Args, 64)
+	if err != nil {
+		return fmt.Errorf("failed to parse caption box opacity: %w", err)
+	}
+	v.Options.Caption.BoxOpacity = opacity
+	return nil
+}
+
+// ExecuteSetCaptionKeyStyle sets the caption key rendering style.
+func ExecuteSetCaptionKeyStyle(c parser.Command, v *VHS) error {
+	v.Options.Caption.KeyStyle = KeyStyle(c.Args)
+	return nil
+}
+
+// ExecuteSetCaptionMarginLeft sets the caption left margin.
+func ExecuteSetCaptionMarginLeft(c parser.Command, v *VHS) error {
+	m, err := strconv.Atoi(c.Args)
+	if err != nil {
+		return fmt.Errorf("failed to parse caption margin left: %w", err)
+	}
+	v.Options.Caption.MarginLeft = m
+	return nil
+}
+
+// ExecuteSetCaptionMarginRight sets the caption right margin.
+func ExecuteSetCaptionMarginRight(c parser.Command, v *VHS) error {
+	m, err := strconv.Atoi(c.Args)
+	if err != nil {
+		return fmt.Errorf("failed to parse caption margin right: %w", err)
+	}
+	v.Options.Caption.MarginRight = m
+	return nil
+}
+
+// ExecuteSetCaptionMarginVertical sets the caption vertical margin.
+func ExecuteSetCaptionMarginVertical(c parser.Command, v *VHS) error {
+	m, err := strconv.Atoi(c.Args)
+	if err != nil {
+		return fmt.Errorf("failed to parse caption margin vertical: %w", err)
+	}
+	v.Options.Caption.MarginVertical = m
+	return nil
+}
+
+// ExecuteSetCaptionAlignment sets the caption alignment.
+func ExecuteSetCaptionAlignment(c parser.Command, v *VHS) error {
+	v.Options.Caption.Alignment = CaptionAlignment(c.Args)
+	return nil
+}
+
+// ExecuteSetCaptionBoxPadding sets the caption box padding.
+func ExecuteSetCaptionBoxPadding(c parser.Command, v *VHS) error {
+	p, err := strconv.Atoi(c.Args)
+	if err != nil {
+		return fmt.Errorf("failed to parse caption box padding: %w", err)
+	}
+	v.Options.Caption.BoxPadding = p
+	return nil
+}
+
+// ExecuteOverlay records a text overlay event at the current frame timestamp.
+func ExecuteOverlay(c parser.Command, v *VHS) error {
+	durationMs := int64(defaultOverlayDurationMs)
+	if c.Options != "" {
+		dur, err := time.ParseDuration(c.Options)
+		if err != nil {
+			return fmt.Errorf("failed to parse overlay duration: %w", err)
+		}
+		durationMs = dur.Milliseconds()
+	}
+
+	frameNum := atomic.LoadInt64(&v.currentFrame)
+	timeMs := frameNum * 1000 / int64(v.Options.Video.Framerate)
+
+	v.OverlayEvents = append(v.OverlayEvents, OverlayEvent{
+		StartMs:    timeMs,
+		DurationMs: durationMs,
+		Text:       c.Args,
+	})
+	return nil
+}
+
+// ExecuteSetOverlayFont sets the overlay font family.
+func ExecuteSetOverlayFont(c parser.Command, v *VHS) error {
+	v.Options.Overlay.Font = c.Args
+	return nil
+}
+
+// ExecuteSetOverlayFontSize sets the overlay font size.
+func ExecuteSetOverlayFontSize(c parser.Command, v *VHS) error {
+	size, err := strconv.Atoi(c.Args)
+	if err != nil {
+		return fmt.Errorf("failed to parse overlay font size: %w", err)
+	}
+	v.Options.Overlay.FontSize = size
+	return nil
+}
+
+// ExecuteSetOverlayFontColor sets the overlay text color.
+func ExecuteSetOverlayFontColor(c parser.Command, v *VHS) error {
+	v.Options.Overlay.FontColor = c.Args
+	return nil
+}
+
+// ExecuteSetOverlayBoxColor sets the overlay background box fill color.
+func ExecuteSetOverlayBoxColor(c parser.Command, v *VHS) error {
+	v.Options.Overlay.BoxColor = c.Args
+	return nil
+}
+
+// ExecuteSetOverlayBoxOpacity sets the overlay box opacity.
+func ExecuteSetOverlayBoxOpacity(c parser.Command, v *VHS) error {
+	opacity, err := strconv.ParseFloat(c.Args, 64)
+	if err != nil {
+		return fmt.Errorf("failed to parse overlay box opacity: %w", err)
+	}
+	v.Options.Overlay.BoxOpacity = opacity
+	return nil
+}
+
+// ExecuteSetOverlayBoxPadding sets the overlay box padding.
+func ExecuteSetOverlayBoxPadding(c parser.Command, v *VHS) error {
+	p, err := strconv.Atoi(c.Args)
+	if err != nil {
+		return fmt.Errorf("failed to parse overlay box padding: %w", err)
+	}
+	v.Options.Overlay.BoxPadding = p
+	return nil
+}
+
+// ExecuteSetOverlayAlignment sets the overlay alignment.
+func ExecuteSetOverlayAlignment(c parser.Command, v *VHS) error {
+	v.Options.Overlay.Alignment = CaptionAlignment(c.Args)
+	return nil
+}
+
+// ExecuteSetOverlayMarginLeft sets the overlay left margin.
+func ExecuteSetOverlayMarginLeft(c parser.Command, v *VHS) error {
+	m, err := strconv.Atoi(c.Args)
+	if err != nil {
+		return fmt.Errorf("failed to parse overlay margin left: %w", err)
+	}
+	v.Options.Overlay.MarginLeft = m
+	return nil
+}
+
+// ExecuteSetOverlayMarginRight sets the overlay right margin.
+func ExecuteSetOverlayMarginRight(c parser.Command, v *VHS) error {
+	m, err := strconv.Atoi(c.Args)
+	if err != nil {
+		return fmt.Errorf("failed to parse overlay margin right: %w", err)
+	}
+	v.Options.Overlay.MarginRight = m
+	return nil
+}
+
+// ExecuteSetOverlayMarginVertical sets the overlay vertical margin.
+func ExecuteSetOverlayMarginVertical(c parser.Command, v *VHS) error {
+	m, err := strconv.Atoi(c.Args)
+	if err != nil {
+		return fmt.Errorf("failed to parse overlay margin vertical: %w", err)
+	}
+	v.Options.Overlay.MarginVertical = m
+	return nil
+}
+
 func getTheme(s string) (Theme, error) {
 	if strings.TrimSpace(s) == "" {
 		return DefaultTheme, nil
diff --git a/command_test.go b/command_test.go
index 19ddc98a..df62f941 100644
--- a/command_test.go
+++ b/command_test.go
@@ -8,12 +8,12 @@ import (
 )
 
 func TestCommand(t *testing.T) {
-	const numberOfCommands = 31
+	const numberOfCommands = 34
 	if len(parser.CommandTypes) != numberOfCommands {
 		t.Errorf("Expected %d commands, got %d", numberOfCommands, len(parser.CommandTypes))
 	}
 
-	const numberOfCommandFuncs = 31
+	const numberOfCommandFuncs = 34
 	if len(CommandFuncs) != numberOfCommandFuncs {
 		t.Errorf("Expected %d commands, got %d", numberOfCommandFuncs, len(CommandFuncs))
 	}
diff --git a/evaluator.go b/evaluator.go
index 5768b854..001428c9 100644
--- a/evaluator.go
+++ b/evaluator.go
@@ -114,6 +114,9 @@ func Evaluate(ctx context.Context, tape string, out io.Writer, opts ...Evaluator
 	ctx, cancel := context.WithCancel(ctx) //nolint:gosec
 	ch := v.Record(ctx)
 
+	// Start key logging
+	v.KeyLogger.Start(&v.currentFrame, v.Options.Video.Framerate)
+
 	// Clean up temporary files at the end.
 	defer func() {
 		if v.Options.Video.Output.Frames != "" {
@@ -182,6 +185,22 @@ func Evaluate(ctx context.Context, tape string, out io.Writer, opts ...Evaluator
 	}
 
 	teardown()
+
+	// Generate caption/overlay ASS file
+	hasCaptions := len(v.KeyLogger.Events()) > 0
+	hasOverlays := len(v.OverlayEvents) > 0
+
+	if hasCaptions || hasOverlays {
+		if err := ensureLibass(); err != nil {
+			return []error{err}
+		}
+		assPath, err := GenerateCaptionFile(v.KeyLogger.Events(), v.OverlayEvents, v.Options.Video, v.Options.Caption, v.Options.Overlay)
+		if err != nil {
+			return []error{err}
+		}
+		v.Options.Video.CaptionFile = assPath
+	}
+
 	if err := v.Render(); err != nil {
 		return []error{err}
 	}
diff --git a/examples/settings/set-caption-alignment-bottom-center.gif b/examples/settings/set-caption-alignment-bottom-center.gif
new file mode 100644
index 00000000..6c3c041c
Binary files /dev/null and b/examples/settings/set-caption-alignment-bottom-center.gif differ
diff --git a/examples/settings/set-caption-alignment-bottom-center.tape b/examples/settings/set-caption-alignment-bottom-center.tape
new file mode 100644
index 00000000..249f71e5
--- /dev/null
+++ b/examples/settings/set-caption-alignment-bottom-center.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-alignment-bottom-center.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionAlignment "bottom-center"
+
+CaptionOn
+
+Type "Where should I go? bottom-center?"
+Sleep 2s
diff --git a/examples/settings/set-caption-alignment-bottom-left.gif b/examples/settings/set-caption-alignment-bottom-left.gif
new file mode 100644
index 00000000..68eb5191
Binary files /dev/null and b/examples/settings/set-caption-alignment-bottom-left.gif differ
diff --git a/examples/settings/set-caption-alignment-bottom-left.tape b/examples/settings/set-caption-alignment-bottom-left.tape
new file mode 100644
index 00000000..8c22502b
--- /dev/null
+++ b/examples/settings/set-caption-alignment-bottom-left.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-alignment-bottom-left.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionAlignment "bottom-left"
+
+CaptionOn
+
+Type "Where should I go? bottom-left?"
+Sleep 2s
diff --git a/examples/settings/set-caption-alignment-bottom-right.gif b/examples/settings/set-caption-alignment-bottom-right.gif
new file mode 100644
index 00000000..67ef5f7a
Binary files /dev/null and b/examples/settings/set-caption-alignment-bottom-right.gif differ
diff --git a/examples/settings/set-caption-alignment-bottom-right.tape b/examples/settings/set-caption-alignment-bottom-right.tape
new file mode 100644
index 00000000..e52fcbd7
--- /dev/null
+++ b/examples/settings/set-caption-alignment-bottom-right.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-alignment-bottom-right.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionAlignment "bottom-right"
+
+CaptionOn
+
+Type "Where should I go? bottom-right?"
+Sleep 2s
diff --git a/examples/settings/set-caption-alignment-middle-center.gif b/examples/settings/set-caption-alignment-middle-center.gif
new file mode 100644
index 00000000..ed41c777
Binary files /dev/null and b/examples/settings/set-caption-alignment-middle-center.gif differ
diff --git a/examples/settings/set-caption-alignment-middle-center.tape b/examples/settings/set-caption-alignment-middle-center.tape
new file mode 100644
index 00000000..cd414b65
--- /dev/null
+++ b/examples/settings/set-caption-alignment-middle-center.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-alignment-middle-center.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionAlignment "middle-center"
+
+CaptionOn
+
+Type "Where should I go? middle-center"
+Sleep 2s
diff --git a/examples/settings/set-caption-alignment-middle-left.gif b/examples/settings/set-caption-alignment-middle-left.gif
new file mode 100644
index 00000000..931f1b22
Binary files /dev/null and b/examples/settings/set-caption-alignment-middle-left.gif differ
diff --git a/examples/settings/set-caption-alignment-middle-left.tape b/examples/settings/set-caption-alignment-middle-left.tape
new file mode 100644
index 00000000..f157fb12
--- /dev/null
+++ b/examples/settings/set-caption-alignment-middle-left.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-alignment-middle-left.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionAlignment "middle-left"
+
+CaptionOn
+
+Type "Where should I go? middle-left"
+Sleep 2s
diff --git a/examples/settings/set-caption-alignment-middle-right.gif b/examples/settings/set-caption-alignment-middle-right.gif
new file mode 100644
index 00000000..f576dbcb
Binary files /dev/null and b/examples/settings/set-caption-alignment-middle-right.gif differ
diff --git a/examples/settings/set-caption-alignment-middle-right.tape b/examples/settings/set-caption-alignment-middle-right.tape
new file mode 100644
index 00000000..527252cb
--- /dev/null
+++ b/examples/settings/set-caption-alignment-middle-right.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-alignment-middle-right.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionAlignment "middle-right"
+
+CaptionOn
+
+Type "Where should I go? middle-right"
+Sleep 2s
diff --git a/examples/settings/set-caption-alignment-top-center.gif b/examples/settings/set-caption-alignment-top-center.gif
new file mode 100644
index 00000000..7926bedf
Binary files /dev/null and b/examples/settings/set-caption-alignment-top-center.gif differ
diff --git a/examples/settings/set-caption-alignment-top-center.tape b/examples/settings/set-caption-alignment-top-center.tape
new file mode 100644
index 00000000..30ea61bb
--- /dev/null
+++ b/examples/settings/set-caption-alignment-top-center.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-alignment-top-center.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionAlignment "top-center"
+
+CaptionOn
+
+Type "Where should I go? top-center?"
+Sleep 2s
diff --git a/examples/settings/set-caption-alignment-top-left.gif b/examples/settings/set-caption-alignment-top-left.gif
new file mode 100644
index 00000000..5bf09737
Binary files /dev/null and b/examples/settings/set-caption-alignment-top-left.gif differ
diff --git a/examples/settings/set-caption-alignment-top-left.tape b/examples/settings/set-caption-alignment-top-left.tape
new file mode 100644
index 00000000..b8d03100
--- /dev/null
+++ b/examples/settings/set-caption-alignment-top-left.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-alignment-top-left.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionAlignment "top-left"
+
+CaptionOn
+
+Type "Where should I go? top-left?"
+Sleep 2s
diff --git a/examples/settings/set-caption-alignment-top-right.gif b/examples/settings/set-caption-alignment-top-right.gif
new file mode 100644
index 00000000..3c40720d
Binary files /dev/null and b/examples/settings/set-caption-alignment-top-right.gif differ
diff --git a/examples/settings/set-caption-alignment-top-right.tape b/examples/settings/set-caption-alignment-top-right.tape
new file mode 100644
index 00000000..2f9d220c
--- /dev/null
+++ b/examples/settings/set-caption-alignment-top-right.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-alignment-top-right.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionAlignment "top-right"
+
+CaptionOn
+
+Type "Where should I go? top-right?"
+Sleep 2s
diff --git a/examples/settings/set-caption-box-color.gif b/examples/settings/set-caption-box-color.gif
new file mode 100644
index 00000000..44f46645
Binary files /dev/null and b/examples/settings/set-caption-box-color.gif differ
diff --git a/examples/settings/set-caption-box-color.tape b/examples/settings/set-caption-box-color.tape
new file mode 100644
index 00000000..94cb60dd
--- /dev/null
+++ b/examples/settings/set-caption-box-color.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-box-color.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionBoxColor "#777777"
+
+CaptionOn
+
+Type "I'm feeling a bit gray today!"
+Sleep 2s
diff --git a/examples/settings/set-caption-box-opacity-0.gif b/examples/settings/set-caption-box-opacity-0.gif
new file mode 100644
index 00000000..56ef3ab9
Binary files /dev/null and b/examples/settings/set-caption-box-opacity-0.gif differ
diff --git a/examples/settings/set-caption-box-opacity-0.tape b/examples/settings/set-caption-box-opacity-0.tape
new file mode 100644
index 00000000..64b5b3db
--- /dev/null
+++ b/examples/settings/set-caption-box-opacity-0.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-box-opacity-0.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionBoxOpacity 0
+
+CaptionOn
+
+Type "I don't want to be boxed in!!"
+Sleep 2s
diff --git a/examples/settings/set-caption-box-opacity-25.gif b/examples/settings/set-caption-box-opacity-25.gif
new file mode 100644
index 00000000..416d6bf7
Binary files /dev/null and b/examples/settings/set-caption-box-opacity-25.gif differ
diff --git a/examples/settings/set-caption-box-opacity-25.tape b/examples/settings/set-caption-box-opacity-25.tape
new file mode 100644
index 00000000..3354d095
--- /dev/null
+++ b/examples/settings/set-caption-box-opacity-25.tape
@@ -0,0 +1,16 @@
+Output examples/settings/set-caption-box-opacity-25.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionBoxOpacity 0.25
+
+CaptionOn
+
+Enter 8
+Type "You can see this text behind the caption box with opacity!"
+Sleep 2s
diff --git a/examples/settings/set-caption-box-opacity-75.gif b/examples/settings/set-caption-box-opacity-75.gif
new file mode 100644
index 00000000..b11cf14c
Binary files /dev/null and b/examples/settings/set-caption-box-opacity-75.gif differ
diff --git a/examples/settings/set-caption-box-opacity-75.tape b/examples/settings/set-caption-box-opacity-75.tape
new file mode 100644
index 00000000..d56b1236
--- /dev/null
+++ b/examples/settings/set-caption-box-opacity-75.tape
@@ -0,0 +1,16 @@
+Output examples/settings/set-caption-box-opacity-75.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionBoxOpacity 0.75
+
+CaptionOn
+
+Enter 8
+Type "You can barely see this when using more opacity!"
+Sleep 2s
diff --git a/examples/settings/set-caption-box-padding-15.gif b/examples/settings/set-caption-box-padding-15.gif
new file mode 100644
index 00000000..5bd6a713
Binary files /dev/null and b/examples/settings/set-caption-box-padding-15.gif differ
diff --git a/examples/settings/set-caption-box-padding-15.tape b/examples/settings/set-caption-box-padding-15.tape
new file mode 100644
index 00000000..a9e81a3c
--- /dev/null
+++ b/examples/settings/set-caption-box-padding-15.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-box-padding-15.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionBoxPadding 15
+
+CaptionOn
+
+Type "I prefer open and more spacious!"
+Sleep 2s
diff --git a/examples/settings/set-caption-box-padding-3.gif b/examples/settings/set-caption-box-padding-3.gif
new file mode 100644
index 00000000..4ce351b7
Binary files /dev/null and b/examples/settings/set-caption-box-padding-3.gif differ
diff --git a/examples/settings/set-caption-box-padding-3.tape b/examples/settings/set-caption-box-padding-3.tape
new file mode 100644
index 00000000..003112c0
--- /dev/null
+++ b/examples/settings/set-caption-box-padding-3.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-box-padding-3.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionBoxPadding 3
+
+CaptionOn
+
+Type "I like it nice and cozy in here!"
+Sleep 2s
diff --git a/examples/settings/set-caption-font-color.gif b/examples/settings/set-caption-font-color.gif
new file mode 100644
index 00000000..2a32b821
Binary files /dev/null and b/examples/settings/set-caption-font-color.gif differ
diff --git a/examples/settings/set-caption-font-color.tape b/examples/settings/set-caption-font-color.tape
new file mode 100644
index 00000000..09c2cc1c
--- /dev/null
+++ b/examples/settings/set-caption-font-color.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-font-color.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionFontColor "#FF893E"
+
+CaptionOn
+
+Type "Look at me! I'm orange now!"
+Sleep 2s
diff --git a/examples/settings/set-caption-font-ibmplex.gif b/examples/settings/set-caption-font-ibmplex.gif
new file mode 100644
index 00000000..01bf1de6
Binary files /dev/null and b/examples/settings/set-caption-font-ibmplex.gif differ
diff --git a/examples/settings/set-caption-font-ibmplex.tape b/examples/settings/set-caption-font-ibmplex.tape
new file mode 100644
index 00000000..0b7a45d6
--- /dev/null
+++ b/examples/settings/set-caption-font-ibmplex.tape
@@ -0,0 +1,19 @@
+Output examples/settings/set-caption-font-ibmplex.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionFont "IBM Plex Mono"
+
+CaptionOn
+
+Type "Change the font"
+Sleep 1.1s
+Left 4
+Sleep 1.1s
+Type "caption "
+Sleep 2s
diff --git a/examples/settings/set-caption-font-jetbrains.gif b/examples/settings/set-caption-font-jetbrains.gif
new file mode 100644
index 00000000..ea1fe1bd
Binary files /dev/null and b/examples/settings/set-caption-font-jetbrains.gif differ
diff --git a/examples/settings/set-caption-font-jetbrains.tape b/examples/settings/set-caption-font-jetbrains.tape
new file mode 100644
index 00000000..27727897
--- /dev/null
+++ b/examples/settings/set-caption-font-jetbrains.tape
@@ -0,0 +1,19 @@
+Output examples/settings/set-caption-font-jetbrains.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionFont "JetBrains Mono"
+
+CaptionOn
+
+Type "Change the font"
+Sleep 1.1s
+Left 4
+Sleep 1.1s
+Type "caption "
+Sleep 2s
diff --git a/examples/settings/set-caption-font-size-18.gif b/examples/settings/set-caption-font-size-18.gif
new file mode 100644
index 00000000..9ab76920
Binary files /dev/null and b/examples/settings/set-caption-font-size-18.gif differ
diff --git a/examples/settings/set-caption-font-size-18.tape b/examples/settings/set-caption-font-size-18.tape
new file mode 100644
index 00000000..3d198f3d
--- /dev/null
+++ b/examples/settings/set-caption-font-size-18.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-font-size-18.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionFontSize 18
+
+CaptionOn
+
+Type "Smaller captions"
+Sleep 2s
diff --git a/examples/settings/set-caption-font-size-26.gif b/examples/settings/set-caption-font-size-26.gif
new file mode 100644
index 00000000..543b5d7e
Binary files /dev/null and b/examples/settings/set-caption-font-size-26.gif differ
diff --git a/examples/settings/set-caption-font-size-26.tape b/examples/settings/set-caption-font-size-26.tape
new file mode 100644
index 00000000..46d4f592
--- /dev/null
+++ b/examples/settings/set-caption-font-size-26.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-font-size-26.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionFontSize 26
+
+CaptionOn
+
+Type "Bigger captions"
+Sleep 2s
diff --git a/examples/settings/set-caption-highlight-color.gif b/examples/settings/set-caption-highlight-color.gif
new file mode 100644
index 00000000..cf048f2e
Binary files /dev/null and b/examples/settings/set-caption-highlight-color.gif differ
diff --git a/examples/settings/set-caption-highlight-color.tape b/examples/settings/set-caption-highlight-color.tape
new file mode 100644
index 00000000..c9600c07
--- /dev/null
+++ b/examples/settings/set-caption-highlight-color.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-highlight-color.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionHighlightColor "#77FFEF"
+
+CaptionOn
+
+Type "Follow me if you can!"
+Sleep 2s
diff --git a/examples/settings/set-caption-inactivity-timer-1200.gif b/examples/settings/set-caption-inactivity-timer-1200.gif
new file mode 100644
index 00000000..352122aa
Binary files /dev/null and b/examples/settings/set-caption-inactivity-timer-1200.gif differ
diff --git a/examples/settings/set-caption-inactivity-timer-1200.tape b/examples/settings/set-caption-inactivity-timer-1200.tape
new file mode 100644
index 00000000..78396472
--- /dev/null
+++ b/examples/settings/set-caption-inactivity-timer-1200.tape
@@ -0,0 +1,23 @@
+Output examples/settings/set-caption-inactivity-timer-1200.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionInactivityTimer 1200ms
+
+CaptionOn
+
+Type "T-minus "
+Sleep 1s
+Type@50ms "3 "
+Sleep 1s
+Type@50ms "2 "
+Sleep 1s
+Type@50ms "1 "
+Sleep 1s
+Type@50ms "Blast off!"
+Sleep 2s
diff --git a/examples/settings/set-caption-inactivity-timer-900.gif b/examples/settings/set-caption-inactivity-timer-900.gif
new file mode 100644
index 00000000..c31175c9
Binary files /dev/null and b/examples/settings/set-caption-inactivity-timer-900.gif differ
diff --git a/examples/settings/set-caption-inactivity-timer-900.tape b/examples/settings/set-caption-inactivity-timer-900.tape
new file mode 100644
index 00000000..d2bfcf4b
--- /dev/null
+++ b/examples/settings/set-caption-inactivity-timer-900.tape
@@ -0,0 +1,23 @@
+Output examples/settings/set-caption-inactivity-timer-900.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionInactivityTimer 900ms
+
+CaptionOn
+
+Type "T-minus "
+Sleep 1s
+Type@50ms "3 "
+Sleep 1s
+Type@50ms "2 "
+Sleep 1s
+Type@50ms "1 "
+Sleep 1s
+Type@50ms "Blast off!"
+Sleep 2s
diff --git a/examples/settings/set-caption-key-style.gif b/examples/settings/set-caption-key-style.gif
new file mode 100644
index 00000000..2a52da81
Binary files /dev/null and b/examples/settings/set-caption-key-style.gif differ
diff --git a/examples/settings/set-caption-key-style.tape b/examples/settings/set-caption-key-style.tape
new file mode 100644
index 00000000..4e066bbb
--- /dev/null
+++ b/examples/settings/set-caption-key-style.tape
@@ -0,0 +1,32 @@
+Output examples/settings/set-caption-key-style.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set Shell zsh
+Set CaptionKeyStyle "vim"
+
+CaptionOn
+
+# Set up line editing to demo control keys
+Hide
+Type "bindkey -e"
+Enter
+Ctrl+l
+Show
+
+Type "echo Goodbye, Planet"
+Sleep 0.5s
+Enter
+Sleep 2s
+Ctrl+p
+Sleep 1.5s
+Ctrl+w
+Sleep 1.1s
+Type "World"
+Enter
+Sleep 2s
diff --git a/examples/settings/set-caption-margin.gif b/examples/settings/set-caption-margin.gif
new file mode 100644
index 00000000..be73e21d
Binary files /dev/null and b/examples/settings/set-caption-margin.gif differ
diff --git a/examples/settings/set-caption-margin.tape b/examples/settings/set-caption-margin.tape
new file mode 100644
index 00000000..6cfd935a
--- /dev/null
+++ b/examples/settings/set-caption-margin.tape
@@ -0,0 +1,17 @@
+Output examples/settings/set-caption-margin.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionAlignment "bottom-right"
+Set CaptionMarginVertical 10
+Set CaptionMarginRight 10
+
+CaptionOn
+
+Type "Margins? Who wants a margin??"
+Sleep 2s
diff --git a/examples/settings/set-caption-max-keys-15.gif b/examples/settings/set-caption-max-keys-15.gif
new file mode 100644
index 00000000..062789ab
Binary files /dev/null and b/examples/settings/set-caption-max-keys-15.gif differ
diff --git a/examples/settings/set-caption-max-keys-15.tape b/examples/settings/set-caption-max-keys-15.tape
new file mode 100644
index 00000000..753370bd
--- /dev/null
+++ b/examples/settings/set-caption-max-keys-15.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-max-keys-15.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionMaxKeys 15
+
+CaptionOn
+
+Type "How much do you want to see?"
+Sleep 2s
diff --git a/examples/settings/set-caption-max-keys-5.gif b/examples/settings/set-caption-max-keys-5.gif
new file mode 100644
index 00000000..9849655a
Binary files /dev/null and b/examples/settings/set-caption-max-keys-5.gif differ
diff --git a/examples/settings/set-caption-max-keys-5.tape b/examples/settings/set-caption-max-keys-5.tape
new file mode 100644
index 00000000..6597565b
--- /dev/null
+++ b/examples/settings/set-caption-max-keys-5.tape
@@ -0,0 +1,15 @@
+Output examples/settings/set-caption-max-keys-5.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set CaptionMaxKeys 5
+
+CaptionOn
+
+Type "How much do you want to see?"
+Sleep 2s
diff --git a/examples/settings/set-caption-toggle-captions.gif b/examples/settings/set-caption-toggle-captions.gif
new file mode 100644
index 00000000..6131f050
Binary files /dev/null and b/examples/settings/set-caption-toggle-captions.gif differ
diff --git a/examples/settings/set-caption-toggle-captions.tape b/examples/settings/set-caption-toggle-captions.tape
new file mode 100644
index 00000000..37a96f9c
--- /dev/null
+++ b/examples/settings/set-caption-toggle-captions.tape
@@ -0,0 +1,31 @@
+Output examples/settings/set-caption-toggle-captions.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set OverlayAlignment "top-right"
+Set OverlayBoxColor "#ffffff"
+Set OverlayBoxOpacity 1
+Set OverlayFontColor "#000000"
+Set OverlayMarginRight 10
+Set OverlayMarginVertical 10
+
+Overlay@100s "Selective Captioning"
+
+CaptionOn
+Type "echo captions are enabled"
+Enter
+Sleep 2s
+
+CaptionOff
+Type "echo captions are disabled"
+Enter
+Sleep 2s
+
+CaptionOn
+Type "echo captions are re-enabled"
+Sleep 2s
diff --git a/examples/settings/set-caption-with-overlay-multiline.gif b/examples/settings/set-caption-with-overlay-multiline.gif
new file mode 100644
index 00000000..76999093
Binary files /dev/null and b/examples/settings/set-caption-with-overlay-multiline.gif differ
diff --git a/examples/settings/set-caption-with-overlay-multiline.tape b/examples/settings/set-caption-with-overlay-multiline.tape
new file mode 100644
index 00000000..9f191066
--- /dev/null
+++ b/examples/settings/set-caption-with-overlay-multiline.tape
@@ -0,0 +1,35 @@
+Output examples/settings/set-caption-with-overlay-multiline.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set Shell zsh
+Set OverlayBoxOpacity 0
+Set OverlayAlignment "top-right"
+
+CaptionOn
+
+# Set up line editing to demo control keys
+Hide
+Type "bindkey -e"
+Enter
+Ctrl+l
+Show
+
+Type "echo Hello, Planet"
+Sleep 0.5s
+Enter
+Sleep 2s
+Overlay@10s "Let's now edit\Nthe prior command\Nto show the world!"
+Sleep 1s
+Ctrl+p
+Sleep 1.5s
+Ctrl+w
+Sleep 1.1s
+Type "World"
+Enter
+Sleep 2s
diff --git a/examples/settings/set-caption-with-overlay-title.gif b/examples/settings/set-caption-with-overlay-title.gif
new file mode 100644
index 00000000..c3dcb708
Binary files /dev/null and b/examples/settings/set-caption-with-overlay-title.gif differ
diff --git a/examples/settings/set-caption-with-overlay-title.tape b/examples/settings/set-caption-with-overlay-title.tape
new file mode 100644
index 00000000..dce0150a
--- /dev/null
+++ b/examples/settings/set-caption-with-overlay-title.tape
@@ -0,0 +1,39 @@
+Output examples/settings/set-caption-with-overlay-title.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set Shell zsh
+Set OverlayAlignment "top-right"
+Set OverlayBoxColor "#ffffff"
+Set OverlayBoxOpacity 1
+Set OverlayFontColor "#000000"
+Set OverlayMarginRight 10
+Set OverlayMarginVertical 10
+
+CaptionOn
+
+# Set up line editing to demo control keys
+Hide
+Type "bindkey -e"
+Enter
+Ctrl+l
+Show
+
+Overlay@100s "Demo of Command Line Editing"
+Type "echo Hello, Planet"
+Sleep 0.5s
+Enter
+Sleep 2s
+Sleep 1s
+Ctrl+p
+Sleep 1.5s
+Ctrl+w
+Sleep 1.1s
+Type "World"
+Enter
+Sleep 2s
diff --git a/examples/settings/set-caption-with-overlay.gif b/examples/settings/set-caption-with-overlay.gif
new file mode 100644
index 00000000..9e16d94f
Binary files /dev/null and b/examples/settings/set-caption-with-overlay.gif differ
diff --git a/examples/settings/set-caption-with-overlay.tape b/examples/settings/set-caption-with-overlay.tape
new file mode 100644
index 00000000..7631ebb6
--- /dev/null
+++ b/examples/settings/set-caption-with-overlay.tape
@@ -0,0 +1,34 @@
+Output examples/settings/set-caption-with-overlay.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set Shell zsh
+Set OverlayAlignment "top-right"
+
+CaptionOn
+
+# Set up line editing to demo control keys
+Hide
+Type "bindkey -e"
+Enter
+Ctrl+l
+Show
+
+Type "echo Hello, Planet"
+Sleep 0.5s
+Enter
+Sleep 2s
+Overlay@10s "Demo of Command Line Editing"
+Sleep 1s
+Ctrl+p
+Sleep 1.5s
+Ctrl+w
+Sleep 1.1s
+Type "World"
+Enter
+Sleep 2s
diff --git a/examples/settings/set-caption.gif b/examples/settings/set-caption.gif
new file mode 100644
index 00000000..0b00dbac
Binary files /dev/null and b/examples/settings/set-caption.gif differ
diff --git a/examples/settings/set-caption.tape b/examples/settings/set-caption.tape
new file mode 100644
index 00000000..a3ce2235
--- /dev/null
+++ b/examples/settings/set-caption.tape
@@ -0,0 +1,31 @@
+Output examples/settings/set-caption.gif
+Set Width 800
+Set Height 225
+Set Padding 30
+
+Set FontSize 18
+Set TypingSpeed 0.1
+Set Theme "Catppuccin Macchiato"
+
+Set Shell zsh
+
+CaptionOn
+
+# Set up line editing to demo control keys
+Hide
+Type "bindkey -e"
+Enter
+Ctrl+l
+Show
+
+Type "echo Hello, Planet"
+Sleep 0.5s
+Enter
+Sleep 2s
+Ctrl+p
+Sleep 1.5s
+Ctrl+w
+Sleep 1.1s
+Type "World"
+Enter
+Sleep 2s
diff --git a/ffmpeg.go b/ffmpeg.go
index d35823af..2eefcfb5 100644
--- a/ffmpeg.go
+++ b/ffmpeg.go
@@ -104,6 +104,30 @@ func calcTermDimensions(style StyleOptions) (int, int) {
 	return width, height
 }
 
+// WithCaptions adds ASS subtitle overlay to ffmpeg filter_complex.
+func (fb *FilterComplexBuilder) WithCaptions(assFilePath string) *FilterComplexBuilder {
+	if assFilePath == "" {
+		return fb
+	}
+	// Escape special characters for ffmpeg filter syntax
+	escaped := strings.ReplaceAll(assFilePath, `\`, `\\`)
+	escaped = strings.ReplaceAll(escaped, `:`, `\:`)
+	escaped = strings.ReplaceAll(escaped, `'`, `'\''`)
+
+	fb.filterComplex.WriteString(";")
+	_, _ = fmt.Fprintf(
+		fb.filterComplex,
+		`
+		[%s]ass='%s'[captioned]
+		`,
+		fb.prevStageName,
+		escaped,
+	)
+	fb.prevStageName = "captioned"
+
+	return fb
+}
+
 // WithWindowBar adds window bar options to ffmepg filter_complex.
 func (fb *FilterComplexBuilder) WithWindowBar(barStream int) *FilterComplexBuilder {
 	if fb.style.WindowBar != "" {
diff --git a/keylogger.go b/keylogger.go
new file mode 100644
index 00000000..74078295
--- /dev/null
+++ b/keylogger.go
@@ -0,0 +1,83 @@
+package main
+
+import (
+	"sync/atomic"
+)
+
+// KeyEvent represents a single key press and timing information.
+type KeyEvent struct {
+	// StartMs is the number of milliseconds relative to the recording start.
+	StartMs int64 `json:"ms"`
+
+	// Key is the string representation of the key pressed.
+	Key string `json:"key"`
+}
+
+// KeyLogger tracks key events during tape execution.
+type KeyLogger struct {
+	events    []KeyEvent
+	paused    bool
+	enabled   bool
+	frame     *int64 // pointer to shared frame counter
+	framerate int
+}
+
+// NewKeyLogger creates a new KeyLogger.
+func NewKeyLogger() *KeyLogger {
+	return &KeyLogger{
+		events: make([]KeyEvent, 0),
+	}
+}
+
+// Start begins key recording.
+//
+// The first parameter is a pointer to the shared frame counter that is being
+// incremented asynchronously by VHS as captures are being made. This is used
+// with the framerate parameter to compute when a key event was made.
+func (l *KeyLogger) Start(frame *int64, framerate int) {
+	l.frame = frame
+	l.framerate = framerate
+}
+
+// Pause suspends key logging until Resume is called.
+func (l *KeyLogger) Pause() {
+	l.paused = true
+}
+
+// Resume enables key logging after Pause is called.
+func (l *KeyLogger) Resume() {
+	l.paused = false
+}
+
+// Enable turns on caption key logging.
+func (l *KeyLogger) Enable() {
+	l.enabled = true
+}
+
+// Disable turns off caption key logging.
+func (l *KeyLogger) Disable() {
+	l.enabled = false
+}
+
+// LogKey records the current key and time at which it occurred with respect
+// to the current frame being captured. If key logging has not been started or
+// if it has been paused, this does nothing.
+func (l *KeyLogger) LogKey(key string) {
+	if l.frame == nil || l.paused || !l.enabled {
+		return
+	}
+
+	frameNum := atomic.LoadInt64(l.frame)
+	timeMs := frameNum * 1000 / int64(l.framerate)
+
+	event := KeyEvent{
+		StartMs: timeMs,
+		Key:     key,
+	}
+	l.events = append(l.events, event)
+}
+
+// Events returns the recorded key events.
+func (l *KeyLogger) Events() []KeyEvent {
+	return l.events
+}
diff --git a/keylogger_test.go b/keylogger_test.go
new file mode 100644
index 00000000..c93019d4
--- /dev/null
+++ b/keylogger_test.go
@@ -0,0 +1,224 @@
+package main
+
+import (
+	"testing"
+)
+
+func TestKeyLogger_Basic(t *testing.T) {
+	l := NewKeyLogger()
+
+	l.LogKey("a")
+	if len(l.events) != 0 {
+		t.Error("Expected no events when recorder has not been started")
+	}
+
+	var frame int64
+	l.Start(&frame, 50) // 50 fps
+	l.Enable()
+
+	l.LogKey("a")
+	frame = 5 // 100ms at 50fps
+	l.LogKey("b")
+
+	// Verify correct number of events
+	if len(l.events) != 2 {
+		t.Errorf("Expected 2 events, got %d", len(l.events))
+	}
+
+	// Verify ordering of events
+	if l.events[0].StartMs >= l.events[1].StartMs {
+		t.Error("Expected first event to have earlier timestamp")
+	}
+
+	// Verify correct timestamps
+	if l.events[0].StartMs != 0 {
+		t.Errorf("Expected first event at 0ms, got %d", l.events[0].StartMs)
+	}
+	if l.events[1].StartMs != 100 {
+		t.Errorf("Expected second event at 100ms, got %d", l.events[1].StartMs)
+	}
+
+	// Verify correct keys
+	if l.events[0].Key != "a" {
+		t.Errorf("Expected first event as 'a', got '%s'", l.events[0].Key)
+	}
+	if l.events[1].Key != "b" {
+		t.Errorf("Expected second event as 'b', got '%s'", l.events[1].Key)
+	}
+
+}
+
+func TestKeyLogger_PauseResume(t *testing.T) {
+	l := NewKeyLogger()
+
+	var frame int64
+	l.Start(&frame, 50) // 50 fps
+	l.Enable()
+
+	l.LogKey("a")
+	l.Pause()
+	l.LogKey("b") // this should not be logged
+	l.Resume()
+	l.LogKey("c")
+
+	// Verify correct number of events
+	if len(l.events) != 2 {
+		t.Errorf("Expected 2 events, got %d", len(l.events))
+	}
+
+	// Verify correct keys
+	if l.events[0].Key != "a" {
+		t.Errorf("Expected first event as 'a', got '%s'", l.events[0].Key)
+	}
+	if l.events[1].Key != "c" {
+		t.Errorf("Expected second event as 'c', got '%s'", l.events[1].Key)
+	}
+}
+
+func TestKeyLogger_Events(t *testing.T) {
+	l := NewKeyLogger()
+
+	var frame int64
+	l.Start(&frame, 50) // 50 fps
+	l.Enable()
+
+	l.LogKey("a")
+	l.LogKey("b")
+	l.LogKey("c")
+
+	events := l.Events()
+	if len(events) != 3 {
+		t.Errorf("Expected 3 events, got %d", len(events))
+	}
+
+	expectedKeys := []string{"a", "b", "c"}
+	for i, expected := range expectedKeys {
+		if events[i].Key != expected {
+			t.Errorf("Event %d: expected key '%s', got '%s'", i, expected, events[i].Key)
+		}
+	}
+}
+
+func TestKeyLogger_EnableDisable(t *testing.T) {
+	l := NewKeyLogger()
+
+	var frame int64
+	l.Start(&frame, 50)
+
+	// Not enabled by default — keys should be ignored
+	l.LogKey("a")
+	if len(l.events) != 0 {
+		t.Errorf("Expected 0 events when disabled, got %d", len(l.events))
+	}
+
+	l.Enable()
+	l.LogKey("b")
+	if len(l.events) != 1 {
+		t.Errorf("Expected 1 event after Enable, got %d", len(l.events))
+	}
+
+	l.Disable()
+	l.LogKey("c")
+	if len(l.events) != 1 {
+		t.Errorf("Expected 1 event after Disable, got %d", len(l.events))
+	}
+
+	l.Enable()
+	l.LogKey("d")
+	if len(l.events) != 2 {
+		t.Errorf("Expected 2 events after re-Enable, got %d", len(l.events))
+	}
+
+	if l.events[0].Key != "b" || l.events[1].Key != "d" {
+		t.Errorf("Expected keys 'b' and 'd', got '%s' and '%s'", l.events[0].Key, l.events[1].Key)
+	}
+}
+
+func TestKeyLogger_PauseAndEnabledInteraction(t *testing.T) {
+	l := NewKeyLogger()
+
+	var frame int64
+	l.Start(&frame, 50)
+	l.Enable()
+
+	l.LogKey("a") // logged
+	l.Pause()
+	l.LogKey("b") // paused — not logged
+	l.Resume()
+	l.LogKey("c") // logged (enabled + not paused)
+
+	// Disable while not paused
+	l.Disable()
+	l.LogKey("d") // disabled — not logged
+
+	// Pause while disabled, then enable
+	l.Pause()
+	l.Enable()
+	l.LogKey("e") // paused — not logged even though enabled
+
+	l.Resume()
+	l.LogKey("f") // enabled + not paused — logged
+
+	if len(l.events) != 3 {
+		t.Errorf("Expected 3 events, got %d", len(l.events))
+	}
+	expectedKeys := []string{"a", "c", "f"}
+	for i, expected := range expectedKeys {
+		if l.events[i].Key != expected {
+			t.Errorf("Event %d: expected key '%s', got '%s'", i, expected, l.events[i].Key)
+		}
+	}
+}
+
+func TestKeyLogger_FrameAlignment(t *testing.T) {
+	tests := []struct {
+		name      string
+		framerate int
+		frames    []int64
+		wantMs    []int64
+	}{
+		{
+			name:      "50fps",
+			framerate: 50,
+			frames:    []int64{0, 25, 50, 100},
+			wantMs:    []int64{0, 500, 1000, 2000},
+		},
+		{
+			name:      "30fps",
+			framerate: 30,
+			frames:    []int64{0, 15, 30, 60},
+			wantMs:    []int64{0, 500, 1000, 2000},
+		},
+		{
+			name:      "60fps",
+			framerate: 60,
+			frames:    []int64{0, 30, 60, 120},
+			wantMs:    []int64{0, 500, 1000, 2000},
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			l := NewKeyLogger()
+			var frame int64
+			l.Start(&frame, tt.framerate)
+			l.Enable()
+
+			for _, f := range tt.frames {
+				frame = f
+				l.LogKey("a")
+			}
+
+			if len(l.events) != len(tt.wantMs) {
+				t.Errorf("Expected %d events, got %d", len(tt.wantMs), len(l.events))
+			}
+
+			for i, want := range tt.wantMs {
+				if l.events[i].StartMs != want {
+					t.Errorf("Event %d: expected %dms, got %dms", i, want, l.events[i].StartMs)
+				}
+			}
+
+		})
+	}
+}
diff --git a/parser/parser.go b/parser/parser.go
index 39a4dcb4..4fbcda69 100644
--- a/parser/parser.go
+++ b/parser/parser.go
@@ -58,6 +58,9 @@ var CommandTypes = []CommandType{
 	token.COPY,
 	token.PASTE,
 	token.ENV,
+	token.OVERLAY,
+	token.CAPTION_ON,
+	token.CAPTION_OFF,
 }
 
 // String returns the string representation of the command.
@@ -185,6 +188,12 @@ func (p *Parser) parseCommand() []Command {
 		return []Command{p.parsePaste()}
 	case token.ENV:
 		return []Command{p.parseEnv()}
+	case token.OVERLAY:
+		return []Command{p.parseOverlay()}
+	case token.CAPTION_ON:
+		return []Command{p.parseCaptionOn()}
+	case token.CAPTION_OFF:
+		return []Command{p.parseCaptionOff()}
 	default:
 		p.errors = append(p.errors, NewError(p.cur, "Invalid command: "+p.cur.Literal))
 		return []Command{{Type: token.ILLEGAL}}
@@ -496,22 +505,10 @@ func (p *Parser) parseSet() Command {
 	case token.MARGIN_FILL:
 		cmd.Args = p.peek.Literal
 		p.nextToken()
-
-		marginFill := p.cur.Literal
-
-		// Check if margin color is a valid hex string
-		if strings.HasPrefix(marginFill, "#") {
-			_, err := strconv.ParseUint(marginFill[1:], 16, 64)
-
-			if err != nil || len(marginFill) != 7 {
-				p.errors = append(
-					p.errors,
-					NewError(
-						p.cur,
-						"\""+marginFill+"\" is not a valid color.",
-					),
-				)
-			}
+		// Margin fill can take non-hex values, so only check on hex values
+		if strings.HasPrefix(p.cur.Literal, "#") && !isValidHexColor(p.cur.Literal) {
+			p.errors = append(p.errors,
+				NewError(p.cur, "\""+p.cur.Literal+"\" is not a valid color."))
 		}
 	case token.CURSOR_BLINK:
 		cmd.Args = p.peek.Literal
@@ -524,6 +521,123 @@ func (p *Parser) parseSet() Command {
 			)
 		}
 
+	case token.CAPTION_KEY_STYLE:
+		cmd.Args = p.peek.Literal
+		p.nextToken()
+		if p.cur.Literal != "vim" && p.cur.Literal != "icon" {
+			p.errors = append(
+				p.errors,
+				NewError(p.cur, "CaptionKeyStyle must be \"vim\" or \"icon\""),
+			)
+		}
+
+	case token.CAPTION_ALIGNMENT:
+		cmd.Args = p.peek.Literal
+		p.nextToken()
+		validAlignments := map[string]bool{
+			"bottom-left": true, "bottom-center": true, "bottom-right": true,
+			"middle-left": true, "middle-center": true, "middle-right": true,
+			"top-left": true, "top-center": true, "top-right": true,
+		}
+		if !validAlignments[p.cur.Literal] {
+			p.errors = append(
+				p.errors,
+				NewError(p.cur, "CaptionAlignment must be one of: bottom-left, bottom-center, bottom-right, middle-left, middle-center, middle-right, top-left, top-center, top-right"),
+			)
+		}
+
+	case token.CAPTION_BOX_OPACITY:
+		cmd.Args = p.peek.Literal
+		p.nextToken()
+		val, err := strconv.ParseFloat(p.cur.Literal, 64)
+		if err != nil || val < 0 || val > 1 {
+			p.errors = append(
+				p.errors,
+				NewError(p.cur, "CaptionBoxOpacity must be a float between 0 and 1"),
+			)
+		}
+
+	case token.CAPTION_FONT:
+		cmd.Args = p.peek.Literal
+		p.nextToken()
+		if len(p.cur.Literal) > 31 {
+			p.errors = append(p.errors,
+				NewError(p.cur, "CaptionFont must not be longer than 31 characters per ASS specification"))
+		}
+
+	case token.CAPTION_HIGHLIGHT_COLOR:
+		cmd.Args = p.peek.Literal
+		p.nextToken()
+		if !isValidHexColor(p.cur.Literal) {
+			p.errors = append(p.errors,
+				NewError(p.cur, "CaptionHighlightColor must be a hex color in #RRGGBB format"))
+		}
+
+	case token.CAPTION_FONT_COLOR:
+		cmd.Args = p.peek.Literal
+		p.nextToken()
+		if !isValidHexColor(p.cur.Literal) {
+			p.errors = append(p.errors,
+				NewError(p.cur, "CaptionFontColor must be a hex color in #RRGGBB format"))
+		}
+
+	case token.CAPTION_BOX_COLOR:
+		cmd.Args = p.peek.Literal
+		p.nextToken()
+		if !isValidHexColor(p.cur.Literal) {
+			p.errors = append(p.errors,
+				NewError(p.cur, "CaptionBoxColor must be a hex color in #RRGGBB format"))
+		}
+
+	case token.CAPTION_INACTIVITY_TIMER:
+		cmd.Args = p.parseTime()
+
+	case token.OVERLAY_FONT:
+		cmd.Args = p.peek.Literal
+		p.nextToken()
+		if len(p.cur.Literal) > 31 {
+			p.errors = append(p.errors,
+				NewError(p.cur, "OverlayFont must not be longer than 31 characters per ASS specification"))
+		}
+
+	case token.OVERLAY_FONT_COLOR:
+		cmd.Args = p.peek.Literal
+		p.nextToken()
+		if !isValidHexColor(p.cur.Literal) {
+			p.errors = append(p.errors,
+				NewError(p.cur, "OverlayFontColor must be a hex color in #RRGGBB format"))
+		}
+
+	case token.OVERLAY_BOX_COLOR:
+		cmd.Args = p.peek.Literal
+		p.nextToken()
+		if !isValidHexColor(p.cur.Literal) {
+			p.errors = append(p.errors,
+				NewError(p.cur, "OverlayBoxColor must be a hex color in #RRGGBB format"))
+		}
+
+	case token.OVERLAY_BOX_OPACITY:
+		cmd.Args = p.peek.Literal
+		p.nextToken()
+		val, err := strconv.ParseFloat(p.cur.Literal, 64)
+		if err != nil || val < 0 || val > 1 {
+			p.errors = append(p.errors,
+				NewError(p.cur, "OverlayBoxOpacity must be a float between 0 and 1"))
+		}
+
+	case token.OVERLAY_ALIGNMENT:
+		cmd.Args = p.peek.Literal
+		p.nextToken()
+		validAlignments := map[string]bool{
+			"bottom-left": true, "bottom-center": true, "bottom-right": true,
+			"middle-left": true, "middle-center": true, "middle-right": true,
+			"top-left": true, "top-center": true, "top-right": true,
+		}
+		if !validAlignments[p.cur.Literal] {
+			p.errors = append(p.errors,
+				NewError(p.cur, "OverlayAlignment must be one of: bottom-left, bottom-center, bottom-right, middle-left, middle-center, middle-right, top-left, top-center, top-right"))
+		}
+
 	default:
 		cmd.Args = p.peek.Literal
 		p.nextToken()
@@ -574,6 +688,20 @@ func (p *Parser) parseShow() Command {
 	return cmd
 }
 
+// parseCaptionOn parses a CaptionOn command.
+//
+//	CaptionOn
+func (p *Parser) parseCaptionOn() Command {
+	return Command{Type: token.CAPTION_ON}
+}
+
+// parseCaptionOff parses a CaptionOff command.
+//
+//	CaptionOff
+func (p *Parser) parseCaptionOff() Command {
+	return Command{Type: token.CAPTION_OFF}
+}
+
 // parseType parses a type command.
 // A type command takes a string to type.
 //
@@ -665,6 +793,31 @@ func (p *Parser) parseEnv() Command {
 	return cmd
 }
 
+// parseOverlay parses an overlay command.
+// An overlay command takes an optional duration and a string to display.
+//
+//	Overlay[@<duration>] "<string>"
+func (p *Parser) parseOverlay() Command {
+	cmd := Command{Type: token.OVERLAY}
+
+	cmd.Options = p.parseSpeed()
+
+	if p.peek.Type != token.STRING {
+		p.errors = append(p.errors, NewError(p.peek, p.cur.Literal+" expects string"))
+	}
+
+	for p.peek.Type == token.STRING {
+		p.nextToken()
+		cmd.Args += p.cur.Literal
+
+		if p.peek.Type == token.STRING {
+			cmd.Args += " "
+		}
+	}
+
+	return cmd
+}
+
 // parseSource parses source command.
 // Source command takes a tape path to include in current tape.
 //
@@ -796,3 +949,12 @@ func isValidWindowBar(w string) bool {
 		w == "Colorful" || w == "ColorfulRight" ||
 		w == "Rings" || w == "RingsRight"
 }
+
+// isValidHexColor checks if a string is a valid #RRGGBB hex color.
+func isValidHexColor(color string) bool {
+	if !strings.HasPrefix(color, "#") || len(color) != 7 {
+		return false
+	}
+	_, err := strconv.ParseUint(color[1:], 16, 64)
+	return err == nil
+}
diff --git a/syntax.go b/syntax.go
index b650cc2c..6131c34e 100644
--- a/syntax.go
+++ b/syntax.go
@@ -59,7 +59,10 @@ func Highlight(c parser.Command, faint bool) string {
 	case token.TYPE:
 		optionsStyle = TimeStyle
 		argsStyle = StringStyle
-	case token.HIDE, token.SHOW:
+	case token.OVERLAY:
+		optionsStyle = TimeStyle
+		argsStyle = StringStyle
+	case token.HIDE, token.SHOW, token.CAPTION_ON, token.CAPTION_OFF:
 		return FaintStyle.Render(c.Type.String())
 	}
 
diff --git a/token/token.go b/token/token.go
index d0d98a1a..080c60e1 100644
--- a/token/token.go
+++ b/token/token.go
@@ -105,70 +105,126 @@ const (
 	WAIT_TIMEOUT    = "WAIT_TIMEOUT"
 	WAIT_PATTERN    = "WAIT_PATTERN"
 	CURSOR_BLINK    = "CURSOR_BLINK"
+
+	CAPTION_ON               = "CAPTION_ON"
+	CAPTION_OFF              = "CAPTION_OFF"
+	CAPTION_FONT             = "CAPTION_FONT"
+	CAPTION_FONT_SIZE        = "CAPTION_FONT_SIZE"
+	CAPTION_MAX_KEYS         = "CAPTION_MAX_KEYS"
+	CAPTION_INACTIVITY_TIMER = "CAPTION_INACTIVITY_TIMER"
+	CAPTION_HIGHLIGHT_COLOR  = "CAPTION_HIGHLIGHT_COLOR"
+	CAPTION_FONT_COLOR       = "CAPTION_FONT_COLOR"
+	CAPTION_BOX_COLOR        = "CAPTION_BOX_COLOR"
+	CAPTION_BOX_OPACITY      = "CAPTION_BOX_OPACITY"
+	CAPTION_KEY_STYLE        = "CAPTION_KEY_STYLE"
+	CAPTION_MARGIN_LEFT      = "CAPTION_MARGIN_LEFT"
+	CAPTION_MARGIN_RIGHT     = "CAPTION_MARGIN_RIGHT"
+	CAPTION_MARGIN_VERTICAL  = "CAPTION_MARGIN_VERTICAL"
+	CAPTION_ALIGNMENT        = "CAPTION_ALIGNMENT"
+	CAPTION_BOX_PADDING      = "CAPTION_BOX_PADDING"
+
+	OVERLAY                 = "OVERLAY"
+	OVERLAY_FONT            = "OVERLAY_FONT"
+	OVERLAY_FONT_SIZE       = "OVERLAY_FONT_SIZE"
+	OVERLAY_FONT_COLOR      = "OVERLAY_FONT_COLOR"
+	OVERLAY_BOX_COLOR       = "OVERLAY_BOX_COLOR"
+	OVERLAY_BOX_OPACITY     = "OVERLAY_BOX_OPACITY"
+	OVERLAY_BOX_PADDING     = "OVERLAY_BOX_PADDING"
+	OVERLAY_ALIGNMENT       = "OVERLAY_ALIGNMENT"
+	OVERLAY_MARGIN_LEFT     = "OVERLAY_MARGIN_LEFT"
+	OVERLAY_MARGIN_RIGHT    = "OVERLAY_MARGIN_RIGHT"
+	OVERLAY_MARGIN_VERTICAL = "OVERLAY_MARGIN_VERTICAL"
 )
 
 // Keywords maps keyword strings to tokens.
 var Keywords = map[string]Type{
-	"em":            EM,
-	"px":            PX,
-	"ms":            MILLISECONDS,
-	"s":             SECONDS,
-	"m":             MINUTES,
-	"Set":           SET,
-	"Sleep":         SLEEP,
-	"Type":          TYPE,
-	"Enter":         ENTER,
-	"Space":         SPACE,
-	"Backspace":     BACKSPACE,
-	"Delete":        DELETE,
-	"Insert":        INSERT,
-	"Ctrl":          CTRL,
-	"Alt":           ALT,
-	"Shift":         SHIFT,
-	"Down":          DOWN,
-	"Left":          LEFT,
-	"Right":         RIGHT,
-	"Up":            UP,
-	"PageUp":        PAGE_UP,
-	"PageDown":      PAGE_DOWN,
-	"ScrollUp":      SCROLL_UP,
-	"ScrollDown":    SCROLL_DOWN,
-	"Tab":           TAB,
-	"Escape":        ESCAPE,
-	"End":           END,
-	"Hide":          HIDE,
-	"Require":       REQUIRE,
-	"Show":          SHOW,
-	"Output":        OUTPUT,
-	"Shell":         SHELL,
-	"FontFamily":    FONT_FAMILY,
-	"MarginFill":    MARGIN_FILL,
-	"Margin":        MARGIN,
-	"WindowBar":     WINDOW_BAR,
-	"WindowBarSize": WINDOW_BAR_SIZE,
-	"BorderRadius":  BORDER_RADIUS,
-	"FontSize":      FONT_SIZE,
-	"Framerate":     FRAMERATE,
-	"Height":        HEIGHT,
-	"LetterSpacing": LETTER_SPACING,
-	"LineHeight":    LINE_HEIGHT,
-	"PlaybackSpeed": PLAYBACK_SPEED,
-	"TypingSpeed":   TYPING_SPEED,
-	"Padding":       PADDING,
-	"Theme":         THEME,
-	"Width":         WIDTH,
-	"LoopOffset":    LOOP_OFFSET,
-	"WaitTimeout":   WAIT_TIMEOUT,
-	"WaitPattern":   WAIT_PATTERN,
-	"Wait":          WAIT,
-	"Source":        SOURCE,
-	"CursorBlink":   CURSOR_BLINK,
-	"true":          BOOLEAN,
-	"false":         BOOLEAN,
-	"Screenshot":    SCREENSHOT,
-	"Copy":          COPY,
-	"Paste":         PASTE,
-	"Env":           ENV,
+	"em":                     EM,
+	"px":                     PX,
+	"ms":                     MILLISECONDS,
+	"s":                      SECONDS,
+	"m":                      MINUTES,
+	"Set":                    SET,
+	"Sleep":                  SLEEP,
+	"Type":                   TYPE,
+	"Enter":                  ENTER,
+	"Space":                  SPACE,
+	"Backspace":              BACKSPACE,
+	"Delete":                 DELETE,
+	"Insert":                 INSERT,
+	"Ctrl":                   CTRL,
+	"Alt":                    ALT,
+	"Shift":                  SHIFT,
+	"Down":                   DOWN,
+	"Left":                   LEFT,
+	"Right":                  RIGHT,
+	"Up":                     UP,
+	"PageUp":                 PAGE_UP,
+	"PageDown":               PAGE_DOWN,
+	"ScrollUp":               SCROLL_UP,
+	"ScrollDown":             SCROLL_DOWN,
+	"Tab":                    TAB,
+	"Escape":                 ESCAPE,
+	"End":                    END,
+	"Hide":                   HIDE,
+	"Require":                REQUIRE,
+	"Show":                   SHOW,
+	"Output":                 OUTPUT,
+	"Shell":                  SHELL,
+	"FontFamily":             FONT_FAMILY,
+	"MarginFill":             MARGIN_FILL,
+	"Margin":                 MARGIN,
+	"WindowBar":              WINDOW_BAR,
+	"WindowBarSize":          WINDOW_BAR_SIZE,
+	"BorderRadius":           BORDER_RADIUS,
+	"FontSize":               FONT_SIZE,
+	"Framerate":              FRAMERATE,
+	"Height":                 HEIGHT,
+	"LetterSpacing":          LETTER_SPACING,
+	"LineHeight":             LINE_HEIGHT,
+	"PlaybackSpeed":          PLAYBACK_SPEED,
+	"TypingSpeed":            TYPING_SPEED,
+	"Padding":                PADDING,
+	"Theme":                  THEME,
+	"Width":                  WIDTH,
+	"LoopOffset":             LOOP_OFFSET,
+	"WaitTimeout":            WAIT_TIMEOUT,
+	"WaitPattern":            WAIT_PATTERN,
+	"Wait":                   WAIT,
+	"Source":                 SOURCE,
+	"CursorBlink":            CURSOR_BLINK,
+	"CaptionOn":              CAPTION_ON,
+	"CaptionOff":             CAPTION_OFF,
+	"CaptionFont":            CAPTION_FONT,
+	"CaptionFontSize":        CAPTION_FONT_SIZE,
+	"CaptionMaxKeys":         CAPTION_MAX_KEYS,
+	"CaptionInactivityTimer": CAPTION_INACTIVITY_TIMER,
+	"CaptionHighlightColor":  CAPTION_HIGHLIGHT_COLOR,
+	"CaptionFontColor":       CAPTION_FONT_COLOR,
+	"CaptionBoxColor":        CAPTION_BOX_COLOR,
+	"CaptionBoxOpacity":      CAPTION_BOX_OPACITY,
+	"CaptionKeyStyle":        CAPTION_KEY_STYLE,
+	"CaptionMarginLeft":      CAPTION_MARGIN_LEFT,
+	"CaptionMarginRight":     CAPTION_MARGIN_RIGHT,
+	"CaptionMarginVertical":  CAPTION_MARGIN_VERTICAL,
+	"CaptionAlignment":       CAPTION_ALIGNMENT,
+	"CaptionBoxPadding":      CAPTION_BOX_PADDING,
+	"Overlay":                OVERLAY,
+	"OverlayFont":            OVERLAY_FONT,
+	"OverlayFontSize":        OVERLAY_FONT_SIZE,
+	"OverlayFontColor":       OVERLAY_FONT_COLOR,
+	"OverlayBoxColor":        OVERLAY_BOX_COLOR,
+	"OverlayBoxOpacity":      OVERLAY_BOX_OPACITY,
+	"OverlayBoxPadding":      OVERLAY_BOX_PADDING,
+	"OverlayAlignment":       OVERLAY_ALIGNMENT,
+	"OverlayMarginLeft":      OVERLAY_MARGIN_LEFT,
+	"OverlayMarginRight":     OVERLAY_MARGIN_RIGHT,
+	"OverlayMarginVertical":  OVERLAY_MARGIN_VERTICAL,
+	"true":                   BOOLEAN,
+	"false":                  BOOLEAN,
+	"Screenshot":             SCREENSHOT,
+	"Copy":                   COPY,
+	"Paste":                  PASTE,
+	"Env":                    ENV,
 }
 
 // IsSetting returns whether a token is a setting.
@@ -177,7 +233,16 @@ func IsSetting(t Type) bool {
 	case SHELL, FONT_FAMILY, FONT_SIZE, LETTER_SPACING, LINE_HEIGHT,
 		FRAMERATE, TYPING_SPEED, THEME, PLAYBACK_SPEED, HEIGHT, WIDTH,
 		PADDING, LOOP_OFFSET, MARGIN_FILL, MARGIN, WINDOW_BAR,
-		WINDOW_BAR_SIZE, BORDER_RADIUS, CURSOR_BLINK, WAIT_TIMEOUT, WAIT_PATTERN:
+		WINDOW_BAR_SIZE, BORDER_RADIUS, CURSOR_BLINK, WAIT_TIMEOUT, WAIT_PATTERN,
+		CAPTION_FONT, CAPTION_FONT_SIZE, CAPTION_MAX_KEYS, CAPTION_INACTIVITY_TIMER,
+		CAPTION_HIGHLIGHT_COLOR, CAPTION_FONT_COLOR, CAPTION_BOX_COLOR,
+		CAPTION_BOX_OPACITY, CAPTION_KEY_STYLE,
+		CAPTION_MARGIN_LEFT, CAPTION_MARGIN_RIGHT, CAPTION_MARGIN_VERTICAL,
+		CAPTION_ALIGNMENT, CAPTION_BOX_PADDING,
+		OVERLAY_FONT, OVERLAY_FONT_SIZE, OVERLAY_FONT_COLOR,
+		OVERLAY_BOX_COLOR, OVERLAY_BOX_OPACITY, OVERLAY_BOX_PADDING,
+		OVERLAY_ALIGNMENT, OVERLAY_MARGIN_LEFT, OVERLAY_MARGIN_RIGHT,
+		OVERLAY_MARGIN_VERTICAL:
 		return true
 	default:
 		return false
@@ -190,7 +255,8 @@ func IsCommand(t Type) bool {
 	case TYPE, SLEEP,
 		UP, DOWN, RIGHT, LEFT, PAGE_UP, PAGE_DOWN, SCROLL_UP, SCROLL_DOWN,
 		ENTER, BACKSPACE, DELETE, TAB,
-		ESCAPE, HOME, INSERT, END, CTRL, SOURCE, SCREENSHOT, COPY, PASTE, WAIT:
+		ESCAPE, HOME, INSERT, END, CTRL, SOURCE, SCREENSHOT, COPY, PASTE, WAIT,
+		OVERLAY, CAPTION_ON, CAPTION_OFF:
 		return true
 	default:
 		return false
diff --git a/vhs.go b/vhs.go
index 235e7b42..bed19614 100644
--- a/vhs.go
+++ b/vhs.go
@@ -12,6 +12,7 @@ import (
 	"regexp"
 	"strings"
 	"sync"
+	"sync/atomic"
 	"time"
 
 	"github.com/go-rod/rod"
@@ -21,18 +22,21 @@ import (
 
 // VHS is the object that controls the setup.
 type VHS struct {
-	Options      *Options
-	Errors       []error
-	Page         *rod.Page
-	browser      *rod.Browser
-	TextCanvas   *rod.Element
-	CursorCanvas *rod.Element
-	mutex        *sync.Mutex
-	started      bool
-	recording    bool
-	tty          *exec.Cmd
-	totalFrames  int
-	close        func() error
+	Options       *Options
+	Errors        []error
+	Page          *rod.Page
+	browser       *rod.Browser
+	TextCanvas    *rod.Element
+	CursorCanvas  *rod.Element
+	mutex         *sync.Mutex
+	started       bool
+	recording     bool
+	tty           *exec.Cmd
+	totalFrames   int
+	currentFrame  int64
+	close         func() error
+	KeyLogger     *KeyLogger
+	OverlayEvents []OverlayEvent
 }
 
 // Options is the set of options for the setup.
@@ -52,6 +56,8 @@ type Options struct {
 	CursorBlink   bool
 	Screenshot    ScreenshotOptions
 	Style         StyleOptions
+	Caption       CaptionOptions
+	Overlay       OverlayOptions
 }
 
 const (
@@ -107,6 +113,8 @@ func DefaultVHSOptions() Options {
 		Screenshot:    screenshot,
 		WaitTimeout:   defaultWaitTimeout,
 		WaitPattern:   defaultWaitPattern,
+		Caption:       DefaultCaptionOptions(),
+		Overlay:       DefaultOverlayOptions(),
 	}
 }
 
@@ -118,6 +126,7 @@ func New() VHS {
 		Options:   &opts,
 		recording: true,
 		mutex:     mu,
+		KeyLogger: NewKeyLogger(),
 	}
 }
 
@@ -359,6 +368,7 @@ func (vhs *VHS) Record(ctx context.Context) <-chan error {
 				}
 
 				counter++
+				atomic.StoreInt64(&vhs.currentFrame, int64(counter))
 				if err := os.WriteFile(
 					filepath.Join(vhs.Options.Video.Input, fmt.Sprintf(cursorFrameFormat, counter)),
 					cursor,
@@ -393,6 +403,7 @@ func (vhs *VHS) ResumeRecording() {
 	defer vhs.mutex.Unlock()
 
 	vhs.recording = true
+	vhs.KeyLogger.Resume()
 }
 
 // PauseRecording indicates to VHS that the recording should be paused.
@@ -401,6 +412,7 @@ func (vhs *VHS) PauseRecording() {
 	defer vhs.mutex.Unlock()
 
 	vhs.recording = false
+	vhs.KeyLogger.Pause()
 }
 
 // ScreenshotNextFrame indicates to VHS that screenshot of next frame must be taken.
diff --git a/video.go b/video.go
index 5548a527..cd6d154e 100644
--- a/video.go
+++ b/video.go
@@ -56,6 +56,7 @@ type VideoOptions struct {
 	Output        VideoOutputs
 	StartingFrame int
 	Style         *StyleOptions
+	CaptionFile   string
 }
 
 const (
@@ -131,6 +132,7 @@ func buildFFopts(opts VideoOptions, targetFile string) []string {
 		WithCorner()
 
 	filterBuilder := NewVideoFilterBuilder(&opts).
+		WithCaptions(opts.CaptionFile).
 		WithWindowBar(streamBuilder.barStream).
 		WithBorderRadius(streamBuilder.cornerStream).
 		WithMarginFill(streamBuilder.marginStream)