📸 Screenshot

Screenshots in PixWit take two steps: first select a region on the screen, then annotate, save, or pin it in the editor window. This page covers the whole flow — from triggering a capture and selecting a region with UI detection, to annotating, copying, and saving.

Hotkeys are customizable

The hotkeys for screenshot/sticky note, recording, AI features, and more can all be changed under Settings → Hotkey Settings; click the × on the right to clear one, which disables it. The wording shown in the UI follows the actual application.

Quick start

• The "Region capture" button in the main window: same as pressing F1; enters screenshot region selection.
• F1 (default screenshot hotkey): opens the editor window once the region is selected; pressing it again while the editor window is open pins the current image as a sticky note right away.
• To skip region selection and capture the whole screen, enable Settings → Screenshot settings → Full screen capture.
• The main window toolbar supports auto-hide at the screen edge: drag it to a screen edge and it snaps there; when the mouse moves away it collapses into a thin strip and pops back out when the mouse returns. Collapsing is not quitting — the screenshot hotkey keeps working as usual.

Triggering and cancelling

How to trigger	Notes
F1 (default)	Enters screenshot region selection; change it under Settings → Hotkey Settings → 📷 Screenshot / Sticky Note
Region capture button in the main window	Same as pressing F1
Single click on the tray icon	In General settings, set the tray left/middle/right single-click action to "Region capture" (middle click by default)

• Once the region is selected, the editor window opens; pressing the screenshot key again while the editor window is open pins the current image as a sticky note right away.
• When a modal dialog is open or the app is busy with another task, the hotkey does not respond.
• With Full screen capture (the hotkey grabs the whole screen directly) enabled, pressing the hotkey immediately grabs the entire screen the mouse is on, copies it to the clipboard automatically, and opens the editor window without entering region selection.
• If you are already in a capture, pressing the screenshot key again cancels that capture.

Region selection

Hints for the currently available actions are shown on screen (top-left corner; you can turn them off under Screenshot settings → Hints and display → Show action hint box):

State	Hint text
Dragging	Drag to select a region, ESC to cancel
Selected	Drag the border to adjust, ESC to cancel
Window/control auto-detected	Tab/scroll wheel to switch, click/Enter to confirm, ESC to cancel
Nothing detected	Drag to select, C to pick color, Tab to switch, ESC to cancel

Mouse and keyboard:

Action	Effect
Left-drag	Create a new selection
Drag an edge / corner	Resize (responds within 12 px of an edge; 8 handles in total)
Drag inside the selection	Move the whole selection
Click outside the selection (free mode)	The nearest edge jumps to the click position and keeps dragging
Right-click	Cancel this capture (no copy, no save)
Double-click	Quick confirm: copy + save to history + close the window
Click a detected window/control	Confirm it directly as the selection
Enter / Space	Confirm as the selection (Space only works once a window/control has been detected)
Enter (when already selected)	Quick confirm: copy + save to history + close
Ctrl+C	Copy and open the editor window
Ctrl+X	Copy and close
Ctrl+S	Save (opens the save dialog) and close
Esc	Cancel the whole capture
Arrow keys	Nudge the selection edge by edge: expands outward by 1 px by default, Shift shrinks inward by 1 px, Ctrl uses a 10 px step

• A selection smaller than 10×10 pixels is treated as invalid and cancelled automatically.
• The size appears next to the selection as a number, e.g. 1920 x 1080; when a ratio or size constraint is active, a marker such as 1:1 or [800×600] is appended.

📷 Image placeholder (Figure 1)

Region selection and handles — ideally show all 8 resize handles, the selection size number 1920 x 1080, the top-left action hint, and the mouse pointer in one shot To add the image: put the file at docs/public/images/screenshot-selection.webp, then delete this tip block and uncomment the next line

Selection mode

Settings → Screenshot settings → Selection settings → Selection mode offers three modes:

Mode	Notes
Free selection (default)	Drag freely, with no ratio or size limits
Fixed ratio	Choose 1:1 (square), 4:3, 9:16 (portrait), 16:9 (landscape)
Fixed size (stamp mode)	Choose 800×600 / 1024×768 / 1280×720 / 1920×1080 / 500×500 / Custom (10–9999 px); in this mode a left click places and confirms the selection box, which can only be moved, not resized

The same group has two more settings that relate directly to selection:

Setting	Default	Notes
Dimming opacity	50	At 0 the unselected area is fully black, at 100 it is fully transparent; the slider takes effect in real time
Allow cross-monitor capture	Off	When off, the selection is limited to the screen the mouse is on; with multiple monitors at different scale factors it is best left off

UI detection mode

Settings → Screenshot settings → Selection settings → Window detection offers only three modes (default "None"):

Mode	Notes
None (no auto-detect) (default)	No automatic detection, best performance
Windows only (fast)	Detects window bounds automatically; fast, recommended for everyday use
Elements only (precise)	Detects individual controls inside a window; most precise, but slower

• While selecting, press Tab to cycle None → Windows only → Elements only → None; a hint such as Detection mode: Windows only appears in the top-right corner (the values are No detection / Windows only / Elements only, and it disappears after about 3 seconds). The last choice is remembered.
• When several windows/controls overlap, use the scroll wheel to cycle through the different levels at the same position (scroll up for an inner, smaller element, scroll down for an outer, larger window). The ladder is ordered by area from inner to outer, and every step switches to a genuinely different size (duplicate levels of the same size are merged, and near-identical levels that differ by only a few pixels are merged into one step). The current step and the total are shown on the detection box's size label (e.g. 301 x 304 · 3/7).
• Besides "the innermost element under the cursor" and "the whole window", the ladder also includes the intermediate containers between them (card / list row / panel / toolbar / sidebar); older windows without an enumerable UI structure have fewer intermediate steps.
• A level picked with the scroll wheel sticks: a later re-detection triggered by a small mouse movement will not push it back to the innermost level. It only returns to the default level when you move the cursor onto a different element (outside the selected box).
• To confirm as the selection: click the detection box, or press Enter / Space.
• Element detection tries MSAA first and automatically falls back to UIA for programs that do not support it. When detection fails or times out, no hint is shown; move the mouse away and back, or use the scroll wheel to change levels, to try again.

There are only three detection modes

Window detection in the current version has only the three modes listed above and no others; the older documentation's description of detection as "automatic window/UI element recognition" is out of date. Please take this section as authoritative.

📷 Image placeholder (Figure 2)

UI detection box and detection mode hint — ideally show a window/control detection box, the top-right Detection mode: Windows only hint, and the multi-level detection box cycled with the scroll wheel in one shot To add the image: put the file at docs/public/images/screenshot-ui-detect.webp, then delete this tip block and uncomment the next line

Screenshot toolbar

Which toolbar buttons appear, and in what order, is fully customizable (Settings → Annotation settings → Toolbar buttons: the left column is "Buttons not shown" and the right column is "Buttons shown on the toolbar"; select a row and use → ← ↑ ↓ to move it, double-click to flip it between the two columns, and click "Restore default order" to go back to the factory arrangement).

The default order from left to right is:

Screenshot/Record toggle, arrow, brush, rectangle, circle, line, polyline, text, numbering, mosaic, undo, trim solid border, Scrolling Screenshot, Text Recognition, scan QR code/barcode, AI dropdown, watermark, close, save, sticky note, copy

The following 5 AI buttons are hidden by default; add them to the toolbar in the settings when needed: Translation, Extract Table, Extract LaTeX formula, LLM interpretation, AI chat.

A few extra notes:

• AI dropdown: a left click on the main button runs the current "default action"; a left click on ▼ opens the menu to choose another feature (Translation also has a submenu of 10 target languages). The "⚙ Set as default action" entry in the menu changes the main button's default feature.
• The watermark button is a checkable toggle: checked shows the watermark, unchecked hides it (the watermark parameters are on the annotation settings page).
• In recording mode the toolbar shows only the three recording controls: region/full screen, video/GIF, and Start Recording.

Scrolling Screenshot

Select a region first, then click Scrolling Screenshot in the toolbar: a left click starts right away with the last remembered mode, while a right click opens a menu to choose Auto / Manual. For full usage (auto/manual mode, status hints, adjustment keys, and the sticky note generated on completion), see Scrolling Screenshot (Chinese).

Color picker

The color picker is enabled by default (Settings → Screenshot settings → Interaction behavior → Enable color picker) and only works in screenshot mode.

• The magnifier is 110×110 px at 8× zoom; it follows the cursor 20 px to the lower right and flips to the other side near a screen edge.
• The info panel is as wide as the magnifier and shows the color swatch, the color value, and the coordinates in that order; the coordinates are logical coordinates within the current screen, and after keyboard nudging the offset is appended, e.g. [+5, -3].

Key	Function
C	Copy the current color to the clipboard and close the screenshot window about 0.2 seconds later (the hint Copied to clipboard! appears below the panel)
Shift (tap)	Cycle the color format: RGB → #HEX → HEX; the choice is remembered
Arrow keys	Nudge the sampling point by 1 pixel
Shift + arrow keys	Nudge by 10 pixels
R	Reset the nudge offset and return to the mouse position
Right-click	Exit the capture

The copied content contains only the color value, not the coordinates, and its format follows the "Color format" setting (default #HEX):

Color format	Copied content
RGB	(255, 128, 64)
#HEX	#FF803F
HEX	FF803F

When the sampling point has been nudged with the keyboard, the magnifier border turns orange as a reminder.

Watermark

The watermark toggle is the watermark button on the screenshot toolbar (off by default); its parameters are under Settings → Annotation settings → Watermark settings:

Setting	Default	Notes
Watermark text:	PixWit	Up to 100 characters
Font size:	24	8–72 px
Color:	Translucent gray	Click the swatch to choose (opacity supported)
Watermark font:	Default (Microsoft YaHei)	Any installed system font
Position:	Bottom right	Top left / top right / bottom left / bottom right / Tiled (45° tilt)
Margin:	20	0–200 px
Tile spacing:	150	Only shown when the position is "Tiled"
Also add date and time	Unchecked	When checked, a time format can be selected
Time format:	yyyy-MM-dd HH:mm:ss	6 formats available

The watermark is previewed in real time during region selection and in the editor window, and is actually burned into the image on output; the settings page itself has no preview image.

📷 Image placeholder (Figure 3)

Watermark effect and settings — ideally show the bottom-right watermark text, the tiled (45° tilt) effect, and the watermark parameters in Annotation settings in one shot To add the image: put the file at docs/public/images/screenshot-watermark.webp, then delete this tip block and uncomment the next line

Saving and notifications

• The save dialog is titled Save screenshot; you can save as PNG / JPEG / WebP / BMP and also export PDF.
• Screenshot save notification (off by default): when checked, a successful save shows the system notification "Screenshot saved successfully / Saved: <file name>".
• Close window after copying is on by default: after clicking "Copy" or pressing Ctrl+C in a screenshot, the window closes.
• Finish the capture automatically after dragging is off by default: when on, releasing the mouse finishes the capture and the editor window is skipped.
• Enable OCR indexing is off by default: when on, screenshots are text-recognized in the background, and you can search Screenshot History by text.

Screenshot settings page at a glance

Settings → 📸 Screenshot settings has 5 groups:

Group	Settings	Default
Selection settings	Dimming opacity, window detection, full screen capture, allow cross-monitor capture, selection mode (+ ratio / size)	50, None, Off, Off, Free selection
Interaction behavior	Enable color picker, close window after copying, double-click to copy the screenshot and close the window, finish the capture automatically after dragging	On, On, On, Off
Hints and display	Screenshot save notification, show action hint box, show line length label	Off, On, Off
Sticky note settings	Hide sticky note border	Off
Image processing	Enable OCR indexing, trim tolerance, trim margin	Off, 30, 5

Changes on the screenshot settings page are only saved after you click Apply or OK at the bottom of Settings.

---

Editing and annotating screenshots

Once the screenshot region is selected, the editor window opens, with all drawing tools on a draggable floating toolbar. The editor window is always on top and has no always-on-top toggle; rotation, format conversion, printing, and similar operations live in the image viewer.

Toolbar buttons

Toolbar buttons can be freely added, removed, and reordered (Settings → Annotation settings → Toolbar buttons). Default order from left to right:

Button	Hover tooltip	Notes
Screenshot/Record toggle	Switch to Record mode	Switches between screenshot and recording; not shown in the editor window
Arrow	Arrow (right-click: no caption/with caption)	Draw arrows, optionally with caption text
Brush	Brush	Freehand drawing
Rectangle	Rectangle (right-click: filled/outline)	Draw rectangles (can be rounded per the "Default corner radius")
Circle	Circle (right-click: filled/outline)	Draw ellipses
Line	Line (right-click: solid/dashed)	Draw straight lines, or drag one into a curve
Polyline	Polyline (right-click: solid/dashed)	Continuous multi-segment polylines
Text	Text	Inline text annotation
Numbering	Numbering (click again to reset; select and type digits to change)	Auto-incrementing number markers
Mosaic	Mosaic	Mosaic masking, also usable to erase text
Undo	Undo (Ctrl+Z)	Undo the previous step
Trim solid border	Trim solid border	Automatically trims the solid border around the image
Scrolling Screenshot	Scrolling Screenshot (right-click: Auto/Manual)	Long screenshots
Text Recognition	Text Recognition (Ctrl+O)	OCR
Scan QR code/barcode	Scan QR code/barcode	Read the contents of a code
AI dropdown	Click the dropdown arrow to choose other features	AI main button + ▼ menu
Watermark	Add/remove watermark	Watermark toggle (checked = shown, unchecked = hidden)
Close	Close (Esc)	Close directly without saving or copying
Save	Save (Ctrl+S)	Save to a file
Sticky Note	Sticky Note (F1)	Pin the current image on the screen
Copy	Copy (Ctrl+C)	Copy to the clipboard; the window stays open

Another 5 AI buttons are hidden by default; add them to the toolbar in the settings when needed: Translation (Ctrl+T), Extract Table (Ctrl+E), Extract LaTeX formula (Ctrl+M), LLM interpretation (Ctrl+I), AI chat (Ctrl+D).

Default hotkeys for the AI features (all changeable and disableable):

AI feature	Default hotkey
Text Recognition (OCR)	Ctrl+O
Extract Table	Ctrl+E
Extract formula (LaTeX)	Ctrl+M
Translation	Ctrl+T
LLM interpretation (smart interpretation)	Ctrl+I
AI chat	Ctrl+D

While the editor window is open, these hotkeys run the corresponding feature on the current image.

Tool right-click menus

Only the following 6 buttons have a right-click menu (the style of the other tools is changed in the "style popup"). The menu is a checked single-choice list, and choosing an entry activates that tool immediately; the choice affects shapes drawn later and also applies to currently selected shapes of the same kind:

Button	Menu items	Remembered	Notes
Rectangle	Outline / Filled	Per tool	Filled mode fills with the line color + its opacity, and there is no outline
Circle	Outline / Filled	Per tool	Same as above
Line	Solid / Dashed	Per tool	Live preview; saving and exporting both draw it dashed
Polyline	Solid / Dashed	Per tool	Same as above
Arrow	No caption / With caption	Remembered	After choosing "With caption", drawing an arrow automatically pops up the caption input box; choosing "No caption" clears the caption of the selected arrow
Scrolling Screenshot	Auto / Manual	Remembered	The chosen mode runs immediately; a normal left click runs the remembered mode

Two common misconceptions

• The middle mouse button does nothing in the editor window: to switch rectangles/circles between filled and outline, use the right-click menu of the rectangle/circle button or the style popup (the middle-click jump to 100% actual size applies only to the image viewer).
• Thickness can be adjusted at any time: after an object is drawn you can keep adjusting its thickness with the scroll wheel at any time; there is no time limit.

📷 Image placeholder (Figure 4)

Screenshot editor toolbar — ideally show the button order, the AI dropdown expanded, the watermark toggle checked, and hover tooltips such as "Rectangle (right-click: filled/outline)" To add the image: put the file at docs/public/images/screenshot-editor-toolbar.webp, then delete this tip block and uncomment the next line

Style popup

Left-clicking any tool button opens that tool's style popup (it does not steal focus, so you can adjust and draw at the same time). Different tools show different controls:

Control	Applies to	Action	Default
Thickness dots	All	The bigger the dot, the thicker the line; scroll wheel ±1	3
8 preset colors + custom color	All	Click a swatch to change the color; "Custom color..." opens the color dialog	—
Opacity dots	Brush, line, polyline, arrow, circle, rectangle; text (with a box)	Scroll wheel ±5 (a checkerboard background shows the actual density)	100 (fully opaque)
Arrow shape	Arrow	Classic arrow / Tapered arrow	Tapered arrow
Mosaic shape	Mosaic	Brush shape / Rectangle shape / Enable text erasing	Brush shape, text erasing off
Numbering style	Numbering	Hollow circle / No circle / Solid circle	Hollow circle (the solid circle is only offered in the screenshot editor window)
Text border	Text	No box / With box	No box

Opacity is presented as "transparency": the value in the UI is transparency — 0% is the most solid and 100% is fully transparent, and each scroll step up makes it more transparent. Regular tools go down to a minimum of 10%; filled rectangles/circles and the text box can go to 0% (leaving only the outline).

Canvas operations

Action	Effect
Left-drag	Draw with the current tool; dragging on empty space moves the editor window
Space (hold) + drag	Hand mode to move the window
Click an annotation	Click the object itself to select it; for filled rectangles/circles, clicking inside selects them; click empty space to deselect
Drag a selected object	Move it as a whole; drag the 8 control points to scale (lines/arrows have 3 points — start/midpoint/end; the brush has none)
Shift + draw line/arrow	Snap the angle to 15°
Shift + draw circle/rectangle	Constrain to a circle / square
Shift + click a preset color	For text with "With box", use that color as the border color (instead of the text color)
Tab	Cycle the selection through annotatable objects (when several objects overlap near the mouse, cycle among those first)
Delete / Backspace	Delete the selected annotation
Ctrl+Z	Undo: cancels text editing while editing text, discards the polyline while drawing one, cancels the current stroke while drawing; otherwise deletes the last annotation / restores the screenshot image (there is no redo)
Scroll wheel	Adjust the current tool's thickness (±1); for numbering, adjusts the radius
Shift + scroll wheel	Adjust opacity (±5)
Filled rectangle/circle + scroll wheel	Adjust the fill transparency directly
Scroll wheel while editing text / caption	Adjust the font size (±2, range 8–160)
Right-click	While drawing a polyline = finish the polyline; otherwise = exit the editor window directly, without saving or copying
Double-click	While drawing a polyline = finish it; while drawing or with a pending text box = cancel; on an arrow = edit or add a caption; otherwise = copy and close (you can turn off "Double-click to copy the screenshot and close the window" in Screenshot settings)
Enter	When no polyline is being drawn, no numbering is being edited, and text has no focus = copy and close

Adjustments show an immediate hint: Thickness: N px, Transparency: N%, Transparency: N% (spotlight N%), Font: N pt. With "Show line length label" enabled, N px is also shown (off by default).

Spotlight: for a filled rectangle/circle, scrolling the transparency to 100% and continuing upward switches from a fill to a "spotlight" density overlay, and the hint becomes Transparency: N% (spotlight N%) — that is, the whole image is dimmed and only the selected area stays clear.

Tool details

Polyline

• Press and drag out one segment, then release to confirm it (a movement that is too small is treated as a stray click and ignored); move the mouse to preview the next segment.
• Finish: double-click / right-click / Enter, or it completes automatically when you switch to another tool; with fewer than 2 vertices the whole polyline is discarded.
• Cancel: Esc (discards the whole polyline). Step back: Backspace / Delete removes the last point, and when only one point is left the whole polyline is cancelled.

Numbering

• Click the canvas to add a number; it increments automatically. Clicking the numbering button again resets the number to 1 and shows the hint Numbering reset to 1.
• Changing a single number: select it and type digits to build the value (applied in real time); Backspace removes a digit, Enter confirms, Esc restores the original value; switching to another object confirms automatically.
• Size: after selecting it, use the scroll wheel to change the radius, or drag a control point.
• Three styles: Hollow circle (default), No circle (the number alone, in a larger font size), Solid circle (solid fill + white text).

Arrow

• Shape: Classic arrow (line + triangular head) or Tapered arrow (solid tapered body); the default is tapered.
• With a caption: after choosing With caption from the right-click menu, drawing an arrow automatically pops up the caption input box; double-clicking the arrow lets you edit or add a caption again. The caption always has a white rounded background with a thin gray border and moves with the arrow.
• Drag the middle to bend it: select a line or arrow and drag the midpoint control point to bend it into an arc; drag the control point back near the midpoint to restore the straight line. A tapered arrow that is too short cannot be bent (the drag falls back to moving the whole shape) and regains the ability once the arrow gets longer.

Text

• With the text tool selected, click the canvas and an inline input box appears after a moment; Esc cancels, Ctrl+Enter confirms, and clicking outside the box also confirms; empty content creates no annotation.
• The default font size is 21 px (adjustable 11–96), and the font comes from the "Text annotation font" setting (follows the system by default).
• Re-editing: click the text itself to enter editing again (note that selecting text shows no control points, so clicking near the four corners may turn into a resize).
• The With box mode adds a background box: Shift + scroll wheel adjusts the box fill transparency (down to 0%, leaving only the border), and Shift + click a preset color / custom color changes the border color.

Mosaic

• The shape can be a brush (free painting) or a rectangle; with "Enable text erasing" checked it is better suited to text areas.

Saving, copying, and closing

Action	Notes
Save (button / Ctrl+S)	First confirms any text and captions being typed, renders all annotations, then saves to a file; the window closes automatically after about 0.1 seconds
Copy (button / Ctrl+C)	Copies to the clipboard and keeps the window open, so you can keep editing
Copy and close (Ctrl+X or Enter)	Copies to the clipboard, writes to history, and closes the window (there is no toolbar button for this)
Close (button / Esc)	Closes directly without saving or copying. Esc is a three-stage cascade: first cancel the edit in progress → then cancel the tool selection → finally close the window
Sticky Note (button)	Pins the result on the screen
Trim solid border	Uses the top-left pixel as the reference color and automatically trims the solid border around the image; if no border is detected it does nothing; on success it goes straight to a sticky note and closes. Tolerance and margin are adjusted under Settings → Screenshot settings → Image processing
Undo	Same as Ctrl+Z

Sticky notes and the clipboard

After a screenshot, click the Sticky Note button in the toolbar, or press Ctrl+Shift+V to pin the clipboard contents, and the image is pinned on the screen as a reference (always on top, resizable, adjustable opacity, croppable). For sticky note window operations, right-click menus, and the full usage of clipboard sticky notes, see Sticky notes and the clipboard (Chinese).