skills.rest
genmedia-labs
genmedia-labsOFFICIAL
Explore all skills by this author →
GitHub

image-outpainting

Extends image canvases beyond original borders using RunComfy edit model endpoints.

12|2|Updated Aug 12, 2026
Search Tags:#image editing#outpainting#ai image#aspect ratio#runcomfy#uncrop#canvas extension
One-click install
npx skills add https://github.com/genmedia-labs/skills --skill image-outpainting-genmedia-labs
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: image-outpainting
Source: https://github.com/genmedia-labs/skills/tree/main/image-outpainting
Command: npx skills add https://github.com/genmedia-labs/skills --skill image-outpainting-genmedia-labs

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @runcomfy/cli.

What problem does it solve? Extending a photo beyond its original frame — uncropping, changing aspect ratio, or filling in what the camera missed — normally requires manual editing in complex tools. This Skill routes outpainting requests to the right RunComfy edit model via the runcomfy CLI, preserving the original content while generating matching surroundings. ## Core Features & Use Cases - Model Routing: Picks between Nano Banana 2 Edit (default, spatial-language driven), GPT Image 2 Edit (multi-reference style matching), FLUX Kontext Pro (maximum preservation), and brand edit endpoints (Seedream, Dreamina, Qwen, FLUX 2) based on the outpaint type. - Aspect-Ratio Conversion: Converts square images to 16:9 banners, portraits to 9:16 for Reels, or 21:9 hero images while keeping the original subject intact. - Uncropping: Extends portraits downward to reveal full body, restores cropped historical photos with reference images, and expands product shots for catalogs. - Use Case: A marketer needs a 16:9 website banner from a 1:1 product photo. The Skill invokes runcomfy run google/nano-banana-2/edit with an outpaint-shaped prompt and aspect_ratio: "16:9", downloading the extended image to a local output directory. ## Quick Start Ask the agent to extend your image to a 16:9 aspect ratio using the runcomfy CLI, providing the source image URL and describing what should appear in the extended areas.

Frequently Asked Questions about image-outpainting

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

FAQPage Schema
How do I outpaint an image to change its aspect ratio?▼

Run `runcomfy run google/nano-banana-2/edit` with a prompt describing the canvas extension and set the `aspect_ratio` parameter explicitly, such as "16:9". Lead the prompt with the canvas change, describe what extends, and end with preservation language to keep the original content intact.

Which model should I use for image outpainting?▼

Nano Banana 2 Edit is the default for prompt-shaped outpaints like aspect-ratio swaps. Use GPT Image 2 Edit when matching a reference style or compositing elements, FLUX Kontext Pro for maximum preservation of the original, and brand endpoints like Seedream or Qwen for brand-locked continuation.

Can I uncrop a portrait to show more of the subject?▼

Yes, use Nano Banana 2 Edit with a prompt like "Extend the canvas downward to show the subject's full upper body" while describing clothing, lighting, and background continuation. End with instructions to keep the face and visible area exactly as in the input.

What should I do when outpainting results have visible seams?▼

Switch to GPT Image 2 Edit with a style reference image, or use FLUX Kontext Pro with maximum-preservation language. For pixel-precise seam handling, use a ComfyUI outpainting workflow hosted on RunComfy, which the CLI cannot reach directly.

Does this skill support mask-driven inpainting or video outpainting?▼

No, mask-driven edits inside the existing canvas are handled by the separate image-inpainting skill, and extending video canvases is handled by the video-outpainting skill. This skill only extends the canvas of still images.

Related Skills

Explore More Design & Creative Skills→
tribes-protocoltribes-protocol
1

zipbox-image

Read images via OpenRouter vision models and generate media through a metered fal.ai API.

Official
Advanced
JiaxingBIJiaxingBI

ckm:banner-design

Generate multi-format HTML/CSS banners with AI-driven visuals and art direction options.

Community
Advanced
ViewWayViewWay

Image Editing

Automate image editing tasks including inpainting, outpainting, background removal, upscaling, restoration, and style transfer.

Community
Intermediate
agentspace-soagentspace-so
31

video-outpainting

Extend video spatial canvas for aspect-ratio changes via RunComfy CLI.

Official
Intermediate