How do I edit images inside a PDF without a separate app?

Two targets: the image or the page

Select an image on the sheet and it becomes the target — the panel shows its pixel size and every tool applies to it. With nothing selected, the target is the current page rendered as an image, which is how you photo-edit a scan or a flattened page.

Page edits flatten: the page is replaced by the edited image, so live text and vector linework would be lost — the panel warns you when that is the case, records an undo entry, and offers a "Revert last page edit" button that restores the pre-edit page. Editing a placed image updates the annotation in place instead, with nothing flattened.

The local toolbox — nothing uploaded

Each tool opens a preview dialog — run it, inspect the result, then Apply or cancel. Three neighbours round out the panel: "Export as PDF" turns the selected image or page into a PDF, "Paste Screenshot" drops the clipboard onto the sheet (Ctrl+V works anywhere on the page), and "Remove Background" opens the free in-browser remover on pdfdepth.com.

• Compress — quality slider (10–95%), JPEG or WebP output, with before/after sizes shown.
• Upscale — 2× or 4× enlargement by high-quality resampling plus sharpening; it is honest about not inventing detail.
• Resize — exact pixel dimensions with aspect-ratio lock and percentage shortcuts.
• Crop — drag a rectangle, free-form or locked to 1:1, 4:3, 16:9, or 3:2.
• Convert Format — PNG, JPEG, or WebP, choosing a background color when flattening transparency into JPEG.
• Remove Watermark — drag boxes over the mark and the regions are inpainted locally.
• Add Text / Watermark — text or a logo image at nine positions plus a tiled pattern, with size, color, and opacity.

Recoloring and retouching

A placed image can also be recolored straight from its hover toolbar: the color editor extracts the image’s dominant colors and lets you remap each one to a new value — a one-click way to push a photo or logo toward brand colors — with an optional background fill for transparent regions. The same editor appears when you cut a region and paste it back as a transparent overlay.

There is no brightness/contrast filter stack here — color work runs through the dominant-color remap or the AI style tools below. The honest split: precise local operations stay local; anything that has to invent pixels goes through the AI section.

AI edits in place

The AI Edit section sends the target to the Gemini image model — it needs your API key in Settings, and the image leaves the device for those calls. The tools cover region edits on a marked area, adding or removing objects and people, replacing one thing with another, whole-image restyling (oil painting, watercolor, cinematic, sketch), clothing and appearance changes, background swap or extension, blending up to three reference images, generative aspect-ratio expansion, and designed text insertion — business card, flyer, poster, invitation, brochure, resume, certificate, letterhead, social post, story, and thumbnail layouts.

Magic Layer sits alongside: it detects the distinct objects in the image — up to a dozen, with labeled boxes you tick or untick — then lifts each chosen object into its own movable image layer, optionally reconstructing the background so moved objects leave a clean plate behind.

Passport photo, quietly included

The panel also hides a passport-photo maker: upload a headshot, drag it under the 2×2-inch crop frame, and set three markers — crown, eyes, chin. A live compliance checklist reports the U.S. requirements it can measure — 600×600 px at 300 dpi, head height 25–35 mm, eye position, headroom, focus, natural color, and a plain light background — and flags the ones you must verify by eye. A whiten-background control with a tolerance slider cleans up the backdrop, and the output downloads as the photo or as a Letter-size print sheet holding nine exact 2×2-inch photos that print true size at 100%.

Doing it in PDFDepth

Right panel → Image Editing Mode. Click a placed image to target it, or leave nothing selected to edit the page as an image. Pick a tool, check the preview, Apply — and the result stays inside the document, ready to export, sign, or share like the rest of the PDF.