Fix typos in image-based PDF slides without regenerating the entire deck.
On-device OCR + Inpainting + Text rendering.
Slides from NotebookLM, Canva, and similar tools look beautiful but are entirely flattened images. To fix a single typo, you have to regenerate the whole deck.
Using on-device OCR to detect text, Neural Engine inpainting to erase it cleanly, and precise rendering to place new text, SlideTouch lets you edit any pixel in a slide.
Apple Vision framework detects text in image slides with bounding boxes. Runs entirely on your Mac.
Original text is erased and the background restored by the Neural Engine, then new text is rendered in place.
Draw a stroke to select any element. Move, reposition, or remove it. Background is seamlessly filled.
All edits are overlay layers. Unlimited undo/redo. Nothing changes until you export.
Full typography control with fonts, sizes, colors, alignment. Plus rectangles, ellipses, arrows, and freeform curves.
Export as PDF, PNG, JPEG, video slideshow, or animated GIF. Original file is never modified.
Drop your image-based PDF into SlideTouch. Pages are rendered at full resolution.
Click any text to edit it. The original is inpainted away and your new text is rendered in place.
Save as a new PDF, image, or video. The original file is never modified.
Sequoia or later
M1 or later required
Recommended for ML inference
For app + ML models