skills.rest
brillianodhiya
brillianodhiyaCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

video-inpainting

Edit regions across video frames on RunComfy using prompt-driven inpainting models.

Updated Jun 11, 2026
Search Tags:#video editing#comfyui#object removal#watermark removal#runcomfy#Wan 2-7#video inpainting
One-click install
npx skills add https://github.com/brillianodhiya/VisionScript --skill video-inpainting-brillianodhiya
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: video-inpainting
Source: https://github.com/brillianodhiya/VisionScript/tree/main/.agents/skills/video-inpainting
Command: npx skills add https://github.com/brillianodhiya/VisionScript --skill video-inpainting-brillianodhiya

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @runcomfy/cli.

What problem does it solve? Removing objects, watermarks, or wires that appear across many video frames is tedious with frame-by-frame manual editing, and choosing the right model for region edits is confusing. This Skill routes video inpainting requests to the right RunComfy endpoint so a single prompt can fix a region consistently across an entire clip. ## Core Features & Use Cases - Model routing: Picks between Wan 2-7 Edit-Video (default prompt-driven region edits), Lucy Edit Restyle (identity-stable region-aware restyle), and Seedream 4-0 Edit-Sequential (frame-stack treatment) based on the edit intent. - Prompting guidance: Provides spatial-language prompting patterns and preservation instructions so edits stay localized without restyling the whole clip. - Escalation path: Points to ComfyUI workflows (LTX 2-3 inpaint, Flux inpainting, SAM2 tracking) when pixel-precise mask propagation is required. - Use Case: Remove a watermark from the bottom-right corner of a clip by running runcomfy run wan-ai/wan-2-7/edit-video with a prompt describing the region and preservation constraints. ## Quick Start Ask the agent to remove the watermark in the bottom-right corner of your video across all frames using the video-inpainting skill.

Frequently Asked Questions about video-inpainting

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

FAQPage Schema
How do I remove a watermark from a video across all frames?▼

Run Wan 2-7 Edit-Video via the runcomfy CLI with a prompt describing the watermark location, such as the bottom-right corner, and instruct it to preserve all other content. For higher quality, use the LTX 2-3 inpaint ComfyUI workflow for mask-driven propagation.

Which model should I use for video inpainting on RunComfy?▼

Wan 2-7 Edit-Video is the default for prompt-driven region edits described in spatial language. Use Lucy Edit Restyle for identity-locked object or outfit swaps, and Seedream 4-0 Edit-Sequential when treating the clip as independently edited frames.

Can I do pixel-precise mask inpainting on video with the runcomfy CLI?▼

The CLI endpoints are prompt-driven and resolve regions from spatial language, not explicit masks. For pixel-precise mask propagation with SAM2 tracking and temporal-aware inpaint, open the hosted ComfyUI workflows such as LTX 2-3 inpaint in the RunComfy cloud.

Why does my video edit change parts of the frame I wanted preserved?▼

Wan 2-7 may restyle frames if the prompt lacks preservation language. Lead with "Preserve all other content exactly" and make one change per call, since compound edits tend to drift across frames.

What are the limitations of frame-by-frame video inpainting?▼

Editing frames independently with Seedream 4-0 Edit-Sequential degrades temporal consistency on long clips and motion-coherent fills. It suits only short, low-frame-rate sequences where a separate tool re-encodes frames to video.

Related Skills

Explore More Design & Creative Skills→
Pixazo-AIPixazo-AI

happy-horse

Generate videos from text, images, or reference clips via the Happy Horse API on Pixazo.

Official
Intermediate
genmedia-labsgenmedia-labs
12

video-edit

Routes video editing requests to RunComfy models for restyle, motion transfer, and outfit swap.

Official
Intermediate
genmedia-labsgenmedia-labs
12

image-inpainting

Performs mask-driven image inpainting via the RunComfy CLI using Z-Image Turbo Inpainting.

Official
Intermediate
genmedia-labsgenmedia-labs
12

video-inpainting

Removes objects, watermarks, and regions across video frames via the RunComfy CLI.

Official
Intermediate