Image annotation
Circle a gel band, arrow a feature on a micrograph, label the lanes. The image annotation editor lets you mark up a photo right where it lives in your notes, and it doesn't touch the original file to do it. Your markup is a separate, re-editable layer you can change or remove at any time.
Lab images carry meaning that the pixels alone do not. A reviewer needs to know which band is the product, which colony you picked, which cell in the field is the one you counted. ResearchOS gives you a drawing surface for exactly that, and it treats your raw image as something you don't ever want corrupted.
The raw image is never modified
This is the part worth understanding before anything else. When you annotate an image, ResearchOS doesn't flatten your shapes into the photo. The original file stays byte-for-byte identical on disk. Your markup is written to a small companion file next to it, called a sidecar, named Images/{filename}.annot.json. A gel photo named gel-day3.png gets a gel-day3.png.annot.json sidecar holding your shapes, and the photo itself is left untouched.
Because the markup lives in its own file, two things follow that matter for trustworthy records. Your annotations stay re-editable, since every shape comes back as a live object you can move, recolor, or delete. And to revert, you delete the sidecar. That returns the image to its original with no quality loss and no trace of the edit baked in.
Opening the editor
There are two ways into the annotation surface, both starting from a thumbnail of the image you want to mark up.
Click a thumbnail to open its image metadata popup (the same popup covered in the editor guide), then press the Annotate button.
Or, on an image strip thumbnail, click the small pencil control to jump straight into the editor.
Either path opens a full-screen editor with the image fit to your viewport and a toolbar across the top. If the image already has annotations, they load back in ready to edit.
The tools
The toolbar holds a focused set of marking tools, each suited to a common lab gesture.
- Arrow and Line: point at a feature or connect two things. The arrow draws a head, and the line is a plain segment.
- Rectangle and Ellipse: box a region of interest or circle a band, a colony, a cell.
- Freehand pen: trace an irregular outline, like the edge of a tissue section.
- Polygon: click vertex by vertex to bound a region with straight edges, then close it by pressing Enter or clicking back on the first point (Escape cancels a polygon in progress).
- Text label: drop a caption directly on the image, for example to name lanes on a gel or stages on a time-course.
Each shape carries its own color and stroke width (text labels use a font size instead), so you can use red for the result and a thinner gray for context. Pick a tool, draw on the image, and the new shape is selected so you can nudge it into place immediately. Double-click a text label to retype it.
Adjusting what you have drawn
Nothing you place is final. Switch to the select tool to drag any shape, and grab the handles on a box or line to resize it. The color and size controls in the toolbar re-edit whichever shape is selected, so you can recolor an existing arrow without redrawing it. A full undo and redo history covers every change, and the delete control (or the Delete key) removes the selected shape.
When you zoom past the fit point, a small navigator appears in the corner with a preview of the whole image and a green box marking the region you are looking at. Click or drag inside it to jump the view to another part of the image without zooming back out.
When the markup looks right, press Save and the editor writes your shapes to the sidecar and closes. Cancel discards everything since you opened the editor. Saving never writes the raw image.
Annotated images render everywhere
Because the markup is a layer, ResearchOS can paint it on top of the photo wherever the photo appears, and it does. Your circled band and labeled lanes show up in the note's markdown preview, in the image strips, and in the results gallery. The overlay even renders down to thumbnails, so a marked-up image is recognizable at a glance without opening it.
The shapes are stored in the image's natural pixel space, so a single saved annotation scales correctly from a full-width figure all the way down to a 64-pixel thumbnail. Annotate it once and it reads right at every size.