Compare commits

...
10 Commits
Author SHA1 Message Date
TZGyn 2e971d5f8d Darken the screen in frozen capture mode
Frozen mode showed the still image at full brightness until a selection
existed, unlike live mode's dark veil. It now draws the same 35% veil over
everything and re-draws the selected area from the snapshot, keeping the
selection bright. The hint text matches the mode.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
2026-09-25 17:43:59 +08:00
TZGyn e4e20c51d3 Add frozen-screen capture on Ctrl+Shift+Cmd+S
A second hotkey snapshots every display the moment it is pressed and
shows that still image as the overlay, so the pointer and hover states
(tooltips, open menus) are captured. Selection is marquee-only — no
window detection — and the result is cropped from the snapshot rather
than recaptured, keeping the frame exactly as it was.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
2026-09-25 17:40:12 +08:00
TZGyn 26700c59f4 Make exported stroke thickness match the preview
Strokes were sized in view points but exported in image pixels, so on
retina captures the saved/copied annotation looked several times thinner.
Stroke widths are now converted into image pixels when drawn and scaled
back for the preview, keeping preview and export identical.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
2026-09-23 17:28:38 +08:00
TZGyn 195c6a9b1f Enable automatic language detection so CJK text is recognized
Vision defaulted to the first listed language and turned Chinese into
Latin garbage (你好世界 became 177WR). automaticallyDetectsLanguage fixes
it; the smoke test now covers Chinese so it cannot regress.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
2026-09-22 10:54:54 +08:00
TZGyn ecac7885dd Keep the captured app focused during selection
The overlay used to activate Kapture, pushing the target app into
background mode so its chrome (title bar, toolbar, tabs) was captured
dimmed. Input no longer needs activation — the event tap handles it — so
the overlay is a non-activating panel shown without taking focus, and the
app keeps its active appearance in the capture.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
2026-09-20 14:46:25 +08:00
TZGyn ef44ef7ac6 Let system dialogs take clicks during capture; hint Esc on the overlay
The event tap consumes every click while selecting, which made permission
prompts unclickable when they appeared mid-capture. Clicks now pass
through when a system dialog (SecurityAgent / System Settings) is
frontmost, and the overlay shows a visible Esc hint as an escape hatch.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
2026-09-12 22:12:48 +08:00
TZGyn 715bb28c1b Fix supersample blur and click-cancellation on micro-drags
The 3x supersample resized the whole capture (blurring content); only the
corner mask is supersampled now, applied through clip(to😷) with the
image drawn once at native scale. The tap's drag branch also clears the
hovered window on sub-4px jitter, cancelling click-to-capture; the
threshold is applied there too.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
2026-09-09 15:26:19 +08:00
TZGyn 07642ca408 Auto-detect window corner radius; anti-alias the mask
Measure the picked window's real corner radius from the capture itself
(the crop's corner pixels are the desktop background; app content starts
where the color departs), correct for arc anti-aliasing, and fall back to
10pt when ambiguous. The clip mask is supersampled 3x for smooth edges.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
2026-09-09 15:20:21 +08:00
TZGyn 924a692f2b Round window-capture corners with a clip instead of masking()
The masking() call could silently fail against ScreenCaptureKit's pixel
format, leaving sharp corners. Clip to a rounded path and redraw instead,
verified headlessly for corner clipping and orientation.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
2026-09-09 15:15:21 +08:00
TZGyn ae054d70b9 Treat any same-app window as popup risk for window capture
Owned popups composite their parent even when the windows do not
overlap, so the intersection test missed DBeaver-style floating dialogs.
Any other window from the same app now routes the pick to the screen
crop.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
2026-09-09 15:09:09 +08:00
7 changed files with 389 additions and 60 deletions
+31 -13
View File
@@ -33,6 +33,7 @@ final class AppDelegate: NSObject, NSApplicationDelegate {
} }
} }
HotkeyManager.shared.onTrigger = { CaptureController.shared.begin() } HotkeyManager.shared.onTrigger = { CaptureController.shared.begin() }
HotkeyManager.shared.onFrozenTrigger = { CaptureController.shared.begin(frozen: true) }
HotkeyManager.shared.register() HotkeyManager.shared.register()
} }
} }
@@ -44,6 +45,9 @@ struct MenuView: View {
Button("Capture & OCR") { CaptureController.shared.begin() } Button("Capture & OCR") { CaptureController.shared.begin() }
.keyboardShortcut("s", modifiers: [.command, .shift]) .keyboardShortcut("s", modifiers: [.command, .shift])
Button("Capture Frozen Screen") { CaptureController.shared.begin(frozen: true) }
.keyboardShortcut("s", modifiers: [.command, .shift, .control])
Button("Extract from Image File…") { ResultWindow.shared.showOpenPanel() } Button("Extract from Image File…") { ResultWindow.shared.showOpenPanel() }
Divider() Divider()
@@ -119,33 +123,47 @@ enum LoginItem {
} }
} }
func runOCRSelfTest() { func renderTestImage(_ text: String, fontSize: CGFloat = 56) -> CGImage? {
let text = "Kapture OCR test 1234, hello world" let size = NSSize(width: 1900, height: 220)
let size = NSSize(width: 1600, height: 220)
guard let rep = NSBitmapImageRep( guard let rep = NSBitmapImageRep(
bitmapDataPlanes: nil, pixelsWide: Int(size.width), pixelsHigh: Int(size.height), bitmapDataPlanes: nil, pixelsWide: Int(size.width), pixelsHigh: Int(size.height),
bitsPerSample: 8, samplesPerPixel: 4, hasAlpha: true, isPlanar: false, bitsPerSample: 8, samplesPerPixel: 4, hasAlpha: true, isPlanar: false,
colorSpaceName: .deviceRGB, bytesPerRow: 0, bitsPerPixel: 0 colorSpaceName: .deviceRGB, bytesPerRow: 0, bitsPerPixel: 0
) else { exit(1) } ) else { return nil }
NSGraphicsContext.saveGraphicsState() NSGraphicsContext.saveGraphicsState()
NSGraphicsContext.current = NSGraphicsContext(bitmapImageRep: rep) NSGraphicsContext.current = NSGraphicsContext(bitmapImageRep: rep)
NSColor.white.setFill() NSColor.white.setFill()
NSRect(origin: .zero, size: size).fill() NSRect(origin: .zero, size: size).fill()
let attrs: [NSAttributedString.Key: Any] = [ let attrs: [NSAttributedString.Key: Any] = [
.font: NSFont.boldSystemFont(ofSize: 56), .font: NSFont.systemFont(ofSize: fontSize),
.foregroundColor: NSColor.black, .foregroundColor: NSColor.black,
] ]
text.draw(at: NSPoint(x: 60, y: 75), withAttributes: attrs) text.draw(at: NSPoint(x: 60, y: 75), withAttributes: attrs)
NSGraphicsContext.restoreGraphicsState() NSGraphicsContext.restoreGraphicsState()
return rep.cgImage
}
guard let cgImage = rep.cgImage else { exit(1) } func runOCRSelfTest() {
OCRService.shared.recognize(cgImage) { recognized in let cases: [(expected: String, contains: String)] = [
print("expected: \(text)") ("Kapture OCR test 1234, hello world", "Kapture OCR test 1234"),
print("got: \(recognized)") ("你好世界,这是一个测试", "你好世界"),
let ok = recognized == text ]
print(ok ? "PASS" : "FAIL") var index = 0
exit(ok ? 0 : 1)
func runNext() {
guard index < cases.count else { exit(0) }
let current = cases[index]
index += 1
guard let image = renderTestImage(current.expected) else { exit(1) }
OCRService.shared.recognize(image) { recognized in
print("case: \(current.expected)")
print("got: \(recognized)")
let ok = recognized.contains(current.contains)
print(ok ? "PASS" : "FAIL")
if !ok { exit(1) }
runNext()
}
} }
runNext()
RunLoop.main.run() RunLoop.main.run()
} }
+226 -23
View File
@@ -76,29 +76,59 @@ final class CaptureController: NSObject {
// MARK: - Session lifecycle // MARK: - Session lifecycle
func begin() { /// Frozen mode: the screen is snapshotted at hotkey press, so the pointer
/// and any hover state (tooltips, open menus) end up in the capture.
private(set) var isFrozenMode = false
private var frozenSnapshots: [ScreenCapture.DisplaySnapshot] = []
func begin(frozen: Bool = false) {
guard !isSelecting else { return } guard !isSelecting else { return }
guard PermissionManager.isGranted else { guard PermissionManager.isGranted else {
PermissionManager.request() PermissionManager.request()
return return
} }
isSelecting = true isSelecting = true
isFrozenMode = frozen
previousApp = NSWorkspace.shared.frontmostApplication previousApp = NSWorkspace.shared.frontmostApplication
// Activation is async on macOS 26; wait a tick so the app is active // With the event tap driving input we must NOT activate: stealing
// before the overlays appear (otherwise clicks pass through). // focus turns the target app into background mode (dimmed chrome),
NSApp.activate(ignoringOtherApps: true) // and that is what would end up in the screenshot. Activate only on
CaptureController.log("begin: activate issued, active=\(NSApp.isActive)") // the fallback path, where panel events are the only input.
let usesTap = AXIsProcessTrusted()
if !usesTap {
NSApp.activate(ignoringOtherApps: true)
}
CaptureController.log("begin: frozen=\(frozen) usesTap=\(usesTap) active=\(NSApp.isActive)")
if frozen {
ScreenCapture.shared.snapshotAllDisplays(includeCursor: true) { [weak self] snapshots in
DispatchQueue.main.async {
guard let self, self.isSelecting else { return }
self.frozenSnapshots = snapshots
CaptureController.log("frozen: captured \(snapshots.count) displays")
self.showOverlays()
}
}
} else {
showOverlays()
}
}
private func showOverlays() {
DispatchQueue.main.asyncAfter(deadline: .now() + 0.05) { [weak self] in DispatchQueue.main.asyncAfter(deadline: .now() + 0.05) { [weak self] in
guard let self, self.isSelecting else { return } guard let self, self.isSelecting else { return }
CaptureController.log("begin: active=\(NSApp.isActive) — showing overlays")
self.overlays = NSScreen.screens.map { screen in self.overlays = NSScreen.screens.map { screen in
let panel = OverlayWindow(screen: screen, controller: self) let panel = OverlayWindow(screen: screen, controller: self)
(panel.contentView as? OverlayView)?.controller = self if let view = panel.contentView as? OverlayView {
view.controller = self
view.frozenImage = self.frozenSnapshots.first(where: { $0.frame == screen.frame })?.image
}
return panel return panel
} }
self.overlays.forEach { $0.makeKeyAndOrderFront(nil) } // orderFrontRegardless: show without taking key/focus.
self.overlays.forEach { $0.orderFrontRegardless() }
self.overlayIDs = self.overlays.map { $0.windowNumber } self.overlayIDs = self.overlays.map { $0.windowNumber }
CaptureController.log("begin: \(self.overlays.count) overlays, ids=\(self.overlayIDs), active=\(NSApp.isActive)") CaptureController.log("begin: \(self.overlays.count) overlays, ids=\(self.overlayIDs) — prior app stays focused")
self.installEventTap() self.installEventTap()
} }
} }
@@ -121,11 +151,30 @@ final class CaptureController: NSObject {
private func finishSelection(_ rect: NSRect) { private func finishSelection(_ rect: NSRect) {
let ids = overlayIDs let ids = overlayIDs
let frozen = frozenSnapshots
let wasFrozen = isFrozenMode
hideOverlays() hideOverlays()
guard rect.width >= 2, rect.height >= 2 else { guard rect.width >= 2, rect.height >= 2 else {
cancelSelection() cancelSelection()
return return
} }
if wasFrozen {
let midpoint = NSPoint(x: rect.midX, y: rect.midY)
let snapshot = frozen.first(where: { $0.frame.contains(midpoint) }) ?? frozen.first
isSelecting = false
restorePreviousApp()
DispatchQueue.main.async { [self] in
if let snapshot, let image = ScreenCapture.cropFrozen(snapshot, to: rect) {
CaptureController.log("frozen: cropped \(Int(rect.width))x\(Int(rect.height)) -> \(image.width)x\(image.height)")
ResultWindow.shared.show(image: image)
} else {
ResultWindow.shared.show(image: ResultWindow.failureImage("Could not crop the frozen screen."))
}
}
return
}
DispatchQueue.main.async { [self] in DispatchQueue.main.async { [self] in
capture(rect, excludingWindowIDs: ids) capture(rect, excludingWindowIDs: ids)
} }
@@ -144,6 +193,8 @@ final class CaptureController: NSObject {
marqueeRect = nil marqueeRect = nil
hoverWindow = nil hoverWindow = nil
marqueeStart = nil marqueeStart = nil
frozenSnapshots = []
isFrozenMode = false
} }
private func restorePreviousApp() { private func restorePreviousApp() {
@@ -262,9 +313,21 @@ final class CaptureController: NSObject {
return nil return nil
} }
/// System dialogs (permission prompts, System Settings) must always
/// receive their clicks even while we're selecting.
private func isSystemDialogFrontmost() -> Bool {
guard let front = NSWorkspace.shared.frontmostApplication else { return false }
let id = front.bundleIdentifier ?? ""
return id.hasPrefix("com.apple.security") || id.hasPrefix("com.apple.SystemSettings") || id == "com.apple.systempreferences"
}
/// Returns the event to pass through, or nil to consume it. /// Returns the event to pass through, or nil to consume it.
func handleTap(_ type: CGEventType, _ event: CGEvent) -> CGEvent? { func handleTap(_ type: CGEventType, _ event: CGEvent) -> CGEvent? {
guard isSelecting else { return event } guard isSelecting else { return event }
if isSystemDialogFrontmost() {
CaptureController.log("tap: system dialog frontmost — passing through \(type)")
return event
}
switch type { switch type {
case .tapDisabledByUserInput, .tapDisabledByTimeout: case .tapDisabledByUserInput, .tapDisabledByTimeout:
@@ -289,6 +352,8 @@ final class CaptureController: NSObject {
} }
let appKitPoint = cgPointToAppKit(NSPoint(x: event.location.x, y: event.location.y)) let appKitPoint = cgPointToAppKit(NSPoint(x: event.location.x, y: event.location.y))
if type == .mouseMoved { if type == .mouseMoved {
// Frozen mode is marquee-only: no window hover detection.
if isFrozenMode { return event }
let window = topmostWindowContaining(appKitPoint) let window = topmostWindowContaining(appKitPoint)
if window?.frame != hoverFrame { if window?.frame != hoverFrame {
hoverWindow = window hoverWindow = window
@@ -301,8 +366,11 @@ final class CaptureController: NSObject {
} }
return event return event
} }
// Drag: grow the marquee // Drag: grow the marquee — but only after real movement, so a
// click's micro-jitter keeps the hovered window eligible.
guard let start = marqueeStart else { return nil } guard let start = marqueeStart else { return nil }
let distance = hypot(appKitPoint.x - start.x, appKitPoint.y - start.y)
guard distance >= 4 else { return nil }
let rect = NSRect( let rect = NSRect(
x: min(start.x, appKitPoint.x), x: min(start.x, appKitPoint.x),
y: min(start.y, appKitPoint.y), y: min(start.y, appKitPoint.y),
@@ -325,12 +393,18 @@ final class CaptureController: NSObject {
if let rect = marqueeRect, dragDistance >= 4, rect.width >= 2, rect.height >= 2 { if let rect = marqueeRect, dragDistance >= 4, rect.width >= 2, rect.height >= 2 {
CaptureController.log("tap: marquee complete \(rect)") CaptureController.log("tap: marquee complete \(rect)")
finishSelection(rect) finishSelection(rect)
} else if isFrozenMode {
CaptureController.log("tap: frozen click without drag — cancel")
cancelSelection()
} else if let hover = hoverWindow { } else if let hover = hoverWindow {
// An overlapping window from the same app means this is an // An overlapping window from the same app means this is an
// owned popup — desktopIndependentWindow would composite the // owned popup — desktopIndependentWindow would composite the
// parent too, so capture the screen crop instead. // parent too, so capture the screen crop instead.
// Owned popups composite their parent even when the windows
// don't overlap, so ANY other same-app window on screen means
// crop (correct region) instead of the group composite.
let ownedRisk = trackedWindows.contains { let ownedRisk = trackedWindows.contains {
$0.id != hover.id && $0.pid == hover.pid && $0.frame.intersects(hover.frame) $0.id != hover.id && $0.pid == hover.pid
} }
CaptureController.log("tap: window pick id=\(hover.id) pid=\(hover.pid) ownedRisk=\(ownedRisk) \(hover.frame)") CaptureController.log("tap: window pick id=\(hover.id) pid=\(hover.pid) ownedRisk=\(ownedRisk) \(hover.frame)")
finishWindowPick(hover.id, hover.frame, ownedRisk: ownedRisk) finishWindowPick(hover.id, hover.frame, ownedRisk: ownedRisk)
@@ -369,6 +443,41 @@ final class ScreenCapture {
private init() {} private init() {}
/// Estimate the window's corner radius (in image pixels) by scanning the
/// top-left corner: the crop's corner pixels are the desktop background;
/// app content starts where the color departs from it.
static func detectCornerRadius(of image: CGImage) -> CGFloat? {
let w = image.width, h = image.height
guard w > 30, h > 30 else { return nil }
var px = [UInt8](repeating: 0, count: w * h * 4)
guard let c = CGContext(data: &px, width: w, height: h, bitsPerComponent: 8, bytesPerRow: w * 4,
space: CGColorSpaceCreateDeviceRGB(),
bitmapInfo: CGImageAlphaInfo.premultipliedLast.rawValue) else { return nil }
c.draw(image, in: CGRect(x: 0, y: 0, width: w, height: h))
func pixel(x: Int, yTop: Int) -> (r: Int, g: Int, b: Int) {
let row = yTop * w * 4 + x * 4
return (Int(px[row]), Int(px[row + 1]), Int(px[row + 2]))
}
let bg = pixel(x: 0, yTop: 0)
func diffFromBg(_ p: (r: Int, g: Int, b: Int)) -> Int {
abs(p.r - bg.r) + abs(p.g - bg.g) + abs(p.b - bg.b)
}
var radii: [CGFloat] = []
for (limit, axis) in [(h / 3, 0), (w / 3, 1)] {
for i in 0..<limit {
let p = axis == 0 ? pixel(x: 0, yTop: i) : pixel(x: i, yTop: 0)
if diffFromBg(p) > 24 {
radii.append(CGFloat(i))
break
}
}
}
guard radii.count == 2 else { return nil }
let r = (radii[0] + radii[1]) / 2
return (r >= 3 && r <= 40) ? r : nil
}
@available(macOS 14.0, *) @available(macOS 14.0, *)
private func crop(_ image: CGImage, to cgRect: CGRect, display: SCDisplay, roundedCorners: Bool = false) -> CGImage? { private func crop(_ image: CGImage, to cgRect: CGRect, display: SCDisplay, roundedCorners: Bool = false) -> CGImage? {
// cgRect is in SCDisplay space (bottom-left origin, Y flipped vs AppKit). // cgRect is in SCDisplay space (bottom-left origin, Y flipped vs AppKit).
@@ -399,29 +508,123 @@ final class ScreenCapture {
// Apply a rounded-corner alpha mask for window captures, matching // Apply a rounded-corner alpha mask for window captures, matching
// macOS window chrome (≈10pt radius, scaled to image pixels). // macOS window chrome (≈10pt radius, scaled to image pixels).
if roundedCorners { if roundedCorners {
let radius = 10 * scaleX // The screen crop contains the app's real rounded corners against
let maskRect = CGRect(x: 0, y: 0, width: result.width, height: result.height) // the desktop — measure the radius from the corner pixels, then
guard let maskCtx = CGContext( // mask with it (supersampled for smooth edges).
let detected = Self.detectCornerRadius(of: result)
let radius = min(max((detected.map { $0 * 1.23 + 2.6 }) ?? (10 * scaleX), 4), 40)
CaptureController.log("corners: detected=\(detected.map { String(format: "%.1f", $0) } ?? "nil") using=\(Int(radius))px")
// Anti-aliased gray mask: supersample 3x, downsample, then draw
// the image through it ONCE at 1x (no content resampling blur).
let ss = 3
let W = result.width * ss, H = result.height * ss
guard let hiMask = CGContext(
data: nil, width: W, height: H,
bitsPerComponent: 8, bytesPerRow: 0,
space: CGColorSpaceCreateDeviceGray(),
bitmapInfo: CGImageAlphaInfo.none.rawValue
) else { return result }
hiMask.setFillColor(CGColor(gray: 1, alpha: 1))
hiMask.fill(CGRect(x: 0, y: 0, width: W, height: H))
hiMask.setFillColor(CGColor(gray: 0, alpha: 1))
let cornerPath = CGMutablePath()
cornerPath.addRect(CGRect(x: 0, y: 0, width: W, height: H))
cornerPath.addRoundedRect(in: CGRect(x: 0, y: 0, width: W, height: H), cornerWidth: radius * CGFloat(ss), cornerHeight: radius * CGFloat(ss), transform: .identity)
hiMask.addPath(cornerPath)
hiMask.fillPath(using: .evenOdd)
guard let hiMaskImg = hiMask.makeImage(),
let downMask = CGContext(
data: nil, width: result.width, height: result.height,
bitsPerComponent: 8, bytesPerRow: 0,
space: CGColorSpaceCreateDeviceGray(),
bitmapInfo: CGImageAlphaInfo.none.rawValue
) else { return result }
downMask.interpolationQuality = .high
downMask.draw(hiMaskImg, in: CGRect(x: 0, y: 0, width: result.width, height: result.height))
guard let mask1x = downMask.makeImage(),
let ctx = CGContext(
data: nil, width: result.width, height: result.height, data: nil, width: result.width, height: result.height,
bitsPerComponent: 8, bytesPerRow: 0, bitsPerComponent: 8, bytesPerRow: 0,
space: CGColorSpaceCreateDeviceRGB(), space: CGColorSpaceCreateDeviceRGB(),
bitmapInfo: CGImageAlphaInfo.premultipliedLast.rawValue bitmapInfo: CGImageAlphaInfo.premultipliedLast.rawValue
) else { return result } ) else { return result }
maskCtx.setFillColor(CGColor(gray: 1, alpha: 1)) ctx.clip(to: CGRect(x: 0, y: 0, width: result.width, height: result.height), mask: mask1x)
maskCtx.fill(maskRect) ctx.draw(result, in: CGRect(x: 0, y: 0, width: result.width, height: result.height))
maskCtx.setBlendMode(.clear) if let masked = ctx.makeImage() {
maskCtx.setFillColor(CGColor(gray: 0, alpha: 1))
maskCtx.addPath(CGPath(roundedRect: maskRect, cornerWidth: radius, cornerHeight: radius, transform: nil))
maskCtx.fillPath()
maskCtx.setBlendMode(.normal)
if let mask = maskCtx.makeImage(),
let masked = result.masking(mask) {
result = masked result = masked
} }
} }
return result return result
} }
/// Screen frame (AppKit global points) + full-resolution image per display.
struct DisplaySnapshot {
let frame: NSRect
let image: CGImage
}
/// Snapshot every display right now. Used by frozen mode: the images are
/// taken before the overlay appears, so the pointer and hover states
/// (tooltips, menus) are baked in.
func snapshotAllDisplays(includeCursor: Bool, completion: @escaping ([DisplaySnapshot]) -> Void) {
if #available(macOS 14.0, *) {
Task {
guard let content = try? await SCShareableContent.current else {
completion([])
return
}
var out: [DisplaySnapshot] = []
for display in content.displays {
guard let screen = NSScreen.screens.first(where: {
(($0.deviceDescription[NSDeviceDescriptionKey("NSScreenNumber")] as? NSNumber)?.uint32Value ?? 0) == display.displayID
}) else { continue }
let config = SCStreamConfiguration()
let scale = screen.backingScaleFactor
config.width = Int(CGFloat(display.width) * scale)
config.height = Int(CGFloat(display.height) * scale)
config.showsCursor = includeCursor
let filter = SCContentFilter(display: display, excludingWindows: [])
if let image = try? await SCScreenshotManager.captureImage(contentFilter: filter, configuration: config) {
out.append(DisplaySnapshot(frame: screen.frame, image: image))
}
}
completion(out)
}
} else {
var out: [DisplaySnapshot] = []
for screen in NSScreen.screens {
let cgRect = CGRect(
x: screen.frame.minX,
y: (NSScreen.screens.first(where: { $0.frame.origin == .zero })?.frame.height ?? screen.frame.height) - screen.frame.maxY,
width: screen.frame.width,
height: screen.frame.height
)
if let image = CGWindowListCreateImage(cgRect, .optionOnScreenOnly, kCGNullWindowID, [.bestResolution, .boundsIgnoreFraming]) {
out.append(DisplaySnapshot(frame: screen.frame, image: image))
}
}
completion(out)
}
}
/// Crop a selection out of a frozen snapshot (top-left pixel space).
static func cropFrozen(_ snapshot: DisplaySnapshot, to rect: NSRect) -> CGImage? {
let scaleX = CGFloat(snapshot.image.width) / snapshot.frame.width
let scaleY = CGFloat(snapshot.image.height) / snapshot.frame.height
guard scaleX > 0, scaleY > 0 else { return nil }
let localX = rect.minX - snapshot.frame.minX
let localTopY = snapshot.frame.maxY - rect.maxY
let px = CGRect(
x: localX * scaleX,
y: localTopY * scaleY,
width: rect.width * scaleX,
height: rect.height * scaleY
)
let safe = px.intersection(CGRect(x: 0, y: 0, width: CGFloat(snapshot.image.width), height: CGFloat(snapshot.image.height))).integral
guard safe.width >= 1, safe.height >= 1 else { return nil }
return snapshot.image.cropping(to: safe)
}
@available(macOS 14.0, *) @available(macOS 14.0, *)
private func displayContaining(_ rect: NSRect, in content: SCShareableContent) -> SCDisplay? { private func displayContaining(_ rect: NSRect, in content: SCShareableContent) -> SCDisplay? {
let mid = NSPoint(x: rect.midX, y: rect.midY) let mid = NSPoint(x: rect.midX, y: rect.midY)
+33 -12
View File
@@ -4,11 +4,14 @@ import AppKit
final class HotkeyManager { final class HotkeyManager {
static let shared = HotkeyManager() static let shared = HotkeyManager()
/// ⌘⇧S — live capture (ScreenCaptureKit at capture time).
var onTrigger: (() -> Void)? var onTrigger: (() -> Void)?
/// ⌃⇧⌘S — frozen capture (snapshot at hotkey press).
var onFrozenTrigger: (() -> Void)?
private var hotKeyRef: EventHotKeyRef? private var hotKeyRefs: [EventHotKeyRef?] = []
private var eventHandlerRef: EventHandlerRef? private var eventHandlerRef: EventHandlerRef?
private var hotKeyID = EventHotKeyID(signature: 0x4B505452, id: 1) private let signature: OSType = 0x4B505452
private init() {} private init() {}
@@ -17,18 +20,36 @@ final class HotkeyManager {
eventClass: OSType(kEventClassKeyboard), eventClass: OSType(kEventClassKeyboard),
eventKind: UInt32(kEventHotKeyPressed) eventKind: UInt32(kEventHotKeyPressed)
) )
InstallEventHandler(GetApplicationEventTarget(), { _, _, _ in InstallEventHandler(GetApplicationEventTarget(), { _, event, _ in
HotkeyManager.shared.onTrigger?() var hotKeyID = EventHotKeyID()
let status = GetEventParameter(
event,
EventParamName(kEventParamDirectObject),
EventParamType(typeEventHotKeyID),
nil,
MemoryLayout<EventHotKeyID>.size,
nil,
&hotKeyID
)
guard status == noErr else { return noErr }
DispatchQueue.main.async {
switch hotKeyID.id {
case 1: HotkeyManager.shared.onTrigger?()
case 2: HotkeyManager.shared.onFrozenTrigger?()
default: break
}
}
return noErr return noErr
}, 1, &eventType, nil, &eventHandlerRef) }, 1, &eventType, nil, &eventHandlerRef)
RegisterEventHotKey( registerHotkey(id: 1, keyCode: UInt32(kVK_ANSI_S), modifiers: UInt32(cmdKey | shiftKey))
UInt32(kVK_ANSI_S), registerHotkey(id: 2, keyCode: UInt32(kVK_ANSI_S), modifiers: UInt32(cmdKey | shiftKey | controlKey))
UInt32(cmdKey | shiftKey), }
hotKeyID,
GetApplicationEventTarget(), private func registerHotkey(id: UInt32, keyCode: UInt32, modifiers: UInt32) {
0, var ref: EventHotKeyRef?
&hotKeyRef let hotKeyID = EventHotKeyID(signature: signature, id: id)
) RegisterEventHotKey(keyCode, modifiers, hotKeyID, GetApplicationEventTarget(), 0, &ref)
hotKeyRefs.append(ref)
} }
} }
+5
View File
@@ -12,6 +12,11 @@ final class OCRService {
request.recognitionLevel = .accurate request.recognitionLevel = .accurate
request.usesLanguageCorrection = true request.usesLanguageCorrection = true
request.recognitionLanguages = ["en-US", "zh-Hans", "zh-Hant", "ja-JP", "ko-KR"] request.recognitionLanguages = ["en-US", "zh-Hans", "zh-Hant", "ja-JP", "ko-KR"]
// Without this, Vision defaults to the first listed language and
// renders CJK as Latin garbage (verified: 你好世界 -> "177WR").
if #available(macOS 13.0, *) {
request.automaticallyDetectsLanguage = true
}
let handler = VNImageRequestHandler(cgImage: cgImage, options: [:]) let handler = VNImageRequestHandler(cgImage: cgImage, options: [:])
do { do {
+83 -3
View File
@@ -6,8 +6,11 @@ final class OverlayWindow: NSPanel {
init(screen: NSScreen, controller: CaptureController) { init(screen: NSScreen, controller: CaptureController) {
let content = OverlayView() let content = OverlayView()
super.init( super.init(
// Non-activating: the overlay must NOT steal focus, so the app
// being captured keeps its active appearance (undimmed chrome).
// Input comes from the CGEventTap, not panel events.
contentRect: screen.frame, contentRect: screen.frame,
styleMask: [.borderless], styleMask: [.borderless, .nonactivatingPanel],
backing: .buffered, backing: .buffered,
defer: false defer: false
) )
@@ -27,6 +30,9 @@ final class OverlayWindow: NSPanel {
final class OverlayView: NSView { final class OverlayView: NSView {
weak var controller: CaptureController? weak var controller: CaptureController?
/// In frozen mode: the still image to display for this screen (already
/// full resolution). Nil = live mode (dim the real screen).
var frozenImage: CGImage?
override var acceptsFirstResponder: Bool { true } override var acceptsFirstResponder: Bool { true }
@@ -53,12 +59,86 @@ final class OverlayView: NSView {
) )
} }
/// Draw the frozen snapshot filling the screen, then replace its centre
/// hole with the live... no: the image IS the content, drawn opaque.
private func drawFrozen(_ image: CGImage) {
guard let ctx = NSGraphicsContext.current?.cgContext else { return }
ctx.saveGState()
ctx.interpolationQuality = .high
// Image is in pixels, view in points — draw into the view bounds.
ctx.draw(image, in: bounds)
ctx.restoreGState()
}
private func drawSizeLabel(for sel: NSRect, in vf: NSRect) {
let label = "\(Int(sel.width)) × \(Int(sel.height))"
let attrs: [NSAttributedString.Key: Any] = [
.font: NSFont.monospacedDigitSystemFont(ofSize: 12, weight: .medium),
.foregroundColor: NSColor.white,
]
let labelSize = label.size(withAttributes: attrs)
let labelRect = NSRect(
x: vf.maxX - labelSize.width - 8,
y: vf.minY - labelSize.height - 10,
width: labelSize.width + 8,
height: labelSize.height + 6
)
NSColor.black.withAlphaComponent(0.6).setFill()
NSBezierPath(roundedRect: labelRect, xRadius: 4, yRadius: 4).fill()
label.draw(at: NSPoint(x: labelRect.minX + 4, y: labelRect.minY + 3), withAttributes: attrs)
}
override func draw(_ dirtyRect: NSRect) { override func draw(_ dirtyRect: NSRect) {
NSColor.black.withAlphaComponent(0.35).setFill() if let frozenImage {
NSBezierPath(rect: bounds).fill() drawFrozen(frozenImage)
} else {
NSColor.black.withAlphaComponent(0.35).setFill()
NSBezierPath(rect: bounds).fill()
}
// Esc hint — the only way out once clicks are consumed by the tap.
let hint = frozenImage != nil
? "Frozen screen · drag to select · Esc to cancel"
: "Drag to select · hover + click a window to capture it · Esc to cancel"
let hintAttrs: [NSAttributedString.Key: Any] = [
.font: NSFont.monospacedSystemFont(ofSize: 12, weight: .medium),
.foregroundColor: NSColor.white.withAlphaComponent(0.85),
]
let hintSize = hint.size(withAttributes: hintAttrs)
let hintRect = NSRect(
x: (bounds.width - hintSize.width) / 2 - 10,
y: bounds.height - 46,
width: hintSize.width + 20,
height: hintSize.height + 12
)
NSColor.black.withAlphaComponent(0.65).setFill()
NSBezierPath(roundedRect: hintRect, xRadius: 8, yRadius: 8).fill()
hint.draw(at: NSPoint(x: hintRect.minX + 10, y: hintRect.minY + 6), withAttributes: hintAttrs)
guard let controller else { return } guard let controller else { return }
// Frozen mode: same look as live mode — a dark veil over everything,
// with the selection showing the frozen pixels undimmed.
if let frozenImage {
NSColor.black.withAlphaComponent(0.35).setFill()
NSBezierPath(rect: bounds).fill()
if let sel = controller.marqueeRect, sel.width > 0, sel.height > 0 {
let vf = viewRect(sel)
NSGraphicsContext.current?.saveGraphicsState()
NSBezierPath(rect: vf).addClip()
drawFrozen(frozenImage)
NSGraphicsContext.current?.restoreGraphicsState()
let border = NSBezierPath(rect: vf)
border.lineWidth = 1.5
NSColor.controlAccentColor.setStroke()
border.stroke()
drawSizeLabel(for: sel, in: vf)
}
return
}
// Highlight the hovered window (rounded, like real window corners) // Highlight the hovered window (rounded, like real window corners)
if let frame = controller.hoverFrame { if let frame = controller.hoverFrame {
let vf = viewRect(frame) let vf = viewRect(frame)
+9 -7
View File
@@ -195,14 +195,15 @@ final class CaptureModel: ObservableObject {
return rep.cgImage ?? image return rep.cgImage ?? image
} }
func addStroke(_ points: [CGPoint]) { /// `widthPx` is the stroke thickness in IMAGE PIXELS (the preview
/// converts the slider value so exports match what was drawn).
func addStroke(_ points: [CGPoint], widthPx: CGFloat) {
guard points.count > 1 else { return } guard points.count > 1 else { return }
if tool == .redact { if tool == .redact {
// Redaction: a filled black bar from start to end, as thick as the // Redaction: a filled black bar from start to end.
// current stroke width (drawn as a rect for clean edges). strokes.append(Stroke(points: points, width: max(widthPx, 12), color: .black, isRedaction: true))
strokes.append(Stroke(points: points, width: max(strokeWidth, 12), color: .black, isRedaction: true))
} else { } else {
strokes.append(Stroke(points: points, width: strokeWidth, color: drawColor)) strokes.append(Stroke(points: points, width: widthPx, color: drawColor))
} }
redoStack = [] redoStack = []
} }
@@ -575,7 +576,8 @@ struct ScreenshotPane: View {
for point in stroke.points.dropFirst() { for point in stroke.points.dropFirst() {
path.addLine(to: CGPoint(x: origin.x + point.x * scale, y: origin.y + point.y * scale)) path.addLine(to: CGPoint(x: origin.x + point.x * scale, y: origin.y + point.y * scale))
} }
context.stroke(path, with: .color(Color(nsColor: stroke.color)), lineWidth: stroke.width) // stroke.width is in image pixels; scale to view points
context.stroke(path, with: .color(Color(nsColor: stroke.color)), lineWidth: stroke.width * scale)
} }
} }
} }
@@ -636,7 +638,7 @@ struct ScreenshotPane: View {
.onEnded { _ in .onEnded { _ in
if model.tool != .select { if model.tool != .select {
if currentStroke.count > 1 { if currentStroke.count > 1 {
model.addStroke(currentStroke) model.addStroke(currentStroke, widthPx: model.strokeWidth / max(scale, 0.01))
} }
currentStroke = [] currentStroke = []
} else { } else {
+2 -2
View File
@@ -108,9 +108,9 @@ cat > "$APP/Contents/Info.plist" <<'PLIST'
<key>CFBundleIconFile</key> <key>CFBundleIconFile</key>
<string>Icon</string> <string>Icon</string>
<key>CFBundleShortVersionString</key> <key>CFBundleShortVersionString</key>
<string>1.8.2</string> <string>1.10.1</string>
<key>CFBundleVersion</key> <key>CFBundleVersion</key>
<string>82</string> <string>101</string>
<key>LSUIElement</key> <key>LSUIElement</key>
<true/> <true/>
<key>NSHighResolutionCapable</key> <key>NSHighResolutionCapable</key>