skills.rest
Loringtonian
LoringtonianCOMMUNITY
Explore all skills by this author →
𝕏@loringtonGitHub

remove-watermark

Removes visible watermarks from images using OpenCV-based inpaint with optional LaMa fallback.

12|2|Updated Jun 8, 2026
Search Tags:#image editing#batch processing#opencv#inpainting#watermark removal#vision qa
One-click install
npx skills add https://github.com/Loringtonian/second-brain-template --skill remove-watermark
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: remove-watermark
Source: https://github.com/Loringtonian/second-brain-template/tree/main/.claude/skills/remove-watermark
Command: npx skills add https://github.com/Loringtonian/second-brain-template --skill remove-watermark

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill removes visible watermarks from images so you can clean up user-owned visuals without manually masking, cloning, or repeatedly retouching the same mark.

Core Features & Use Cases

  • Fast local inpainting: Uses deterministic OpenCV-based removal for sub-second cleanup on common visible marks.
  • Vision-guided detection: Leverages a vision model to locate the watermark accurately, even when its position shifts across aspect ratios.
  • Edge-case handling: Switches to a higher-quality reconstruction path when the mark sits on a high-contrast boundary to avoid smearing.
  • Batch-friendly workflow: Supports single images or folders, writes non-destructive _clean outputs, and preserves before/after QA crops.
  • Use Case: Remove the Gemini sparkle from a photo, clear a logo from a product shot, or process a folder of images with consistent QA checks.

Quick Start

Use the remove-watermark skill to clean the attached image or folder of images and return the non-destructive _clean results.

Frequently Asked Questions about remove-watermark

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I remove visible watermarks from a batch of images?▼

Watermark removal from a batch of images is handled by processing folders directly. The workflow applies vision-guided detection, local OpenCV inpainting, and outputs non-destructive `_clean` files for every photo.

What is the best way to automatically erase a logo from a product shot?▼

Erasing a logo from a product shot is best handled by a vision-guided locate-and-QA loop. This approach normalizes region targeting, applies local OpenCV inpainting, and switches to a LaMa fallback for edge cases.

Does OpenCV inpainting work for watermarks on high-contrast boundaries?▼

OpenCV inpainting for watermarks on high-contrast boundaries can cause smearing. The workflow detects these edge cases and switches to a higher-quality LaMa reconstruction path to avoid artifacts.

How does vision-based QA work for image cleanup?▼

Vision-based QA for image cleanup works by accurately locating watermarks across varying aspect ratios, verifying the inpainted result, and preserving before-and-after QA crops for visual review.

Can I clean up user-owned photos without altering the original files?▼

Yes, you can clean up user-owned photos without altering the originals. The process writes non-destructive `_clean` output files and preserves before/after QA crops, leaving source images untouched.

Related Skills

Explore More Design & Creative Skills→
SoufianouassifSoufianouassif

photo-editor

Resize, crop, and enhance images programmatically with Pillow and OpenCV.

Community
Intermediate
kk20300113-pngkk20300113-png

design-review

Detect and fix visual design inconsistencies in web interfaces with atomic commits.

Community
Advanced
phuhao00phuhao00
12

image-edit-agent

Performs automated image manipulation including inpainting, outpainting, and watermark removal from natural language instructions.

Community
Advanced
GitnappGitnapp

segment-anything-model

Segment objects in images using point prompts, boxes, or masks.

Community
Advanced