Omasnap

Omasnap

Native Wayland screenshot annotator

Description

Annotating a screenshot on Wayland usually means grim plus slurp and then dragging the result into a separate editor — three or four steps — and the graphical screenshot tools tend to photograph their own editor window in the process. Omasnap rebuilds that path: it captures the focused monitor first, then maps an exclusive layer-shell surface for the editor, so the editor never lands in its own screenshot, and on scaled displays what you get is the monitor's native pixels rather than an upscaled copy.

Annotations are not burned into the bitmap. Arrows, boxes, text and redactions stay as vector layers you can move, resize and reorder afterwards. Drag a layer past the edge of the shot and the canvas grows to meet it, with the new strip rendered as a shadowed window-grey card, a gradient backdrop, or left transparent. The cut tool is the quiet favourite: drag across a band of the image and that band disappears with the gap closed up and the annotations shifted to follow, which makes trimming dead space out of a tall capture painless. Written in C++ for Omarchy and Hyprland under the MIT licence.

Features



Three capture modes: freeform region, window and full monitor, where window capture crops the monitor frame so overlapping windows stay visible instead of triggering a second clean recapture.

Vector annotation layers: arrows, straight lines, smoothed freehand, highlighter strokes, hollow or filled rectangles and ellipses, numbered markers and editable text, all selectable, movable, resizable and recolourable with full undo history.

Growing canvas: dragging a layer off the edge expands the image, cycling between a framed strip, a tight overflow fit and the original size, with gradient, window-grey or transparent backdrops.

Cut and redact: drag across a band to delete it and collapse the gap, with a live preview and a dashed seam marker while you drag, and redact sensitive regions with either opaque fills or a randomised non-spatial mosaic.

OCR and pinning: one click runs text recognition on the whole image or a dragged region, showing the text beside the shot and copying it to the clipboard, and a finished capture can be pinned as an always-on-top surface visible on every workspace.

Crash resistant: working documents live under the runtime directory as the source image plus a JSON operation log, so undo history survives a crash or a reopen.