background-removal

Remove backgrounds from images using BiRefNet via the inference.sh CLI.

Updated Jul 28, 2026
One-click install
npx skills add https://github.com/Aadi-110i/PEP-PROJECT --skill background-removal-aadi-110i
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: background-removal
Source: https://github.com/Aadi-110i/PEP-PROJECT/tree/main/skills/background-removal
Command: npx skills add https://github.com/Aadi-110i/PEP-PROJECT --skill background-removal-aadi-110i

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Removing backgrounds from images manually requires photo editing software and tedious masking work. This Skill automates background removal through the inference.sh CLI, producing transparent PNGs from any image URL. ## Core Features & Use Cases - BiRefNet Background Removal: Run the BiRefNet model via belt app run infsh/birefnet to cut out subjects and return transparent PNGs. - Prompt-Based Editing with Reve: Use the falai/reve app to remove or replace backgrounds with natural language prompts, such as changing a backdrop to a beach scene. - Generation-to-Edit Workflows: Chain image generation (e.g., FLUX) with background removal in a single pipeline. - Use Case: An e-commerce seller needs clean product photos. Pass each product image URL to the BiRefNet app and receive transparent PNGs ready for storefront listings. ## Quick Start Ask the assistant to remove the background from your image URL using the background-removal skill, and it will run the BiRefNet app through the belt CLI and return a transparent PNG.

Frequently Asked Questions about background-removal

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

FAQPage Schema
How do I remove a background from an image with inference.sh?

Run belt app run infsh/birefnet with an input JSON containing your image_url. The BiRefNet model processes the image and returns a PNG with a transparent background.

How to change an image background instead of removing it?

Use the falai/reve app with a prompt describing the new background, such as changing it to a beach scene. Pass the prompt and image_url in the input JSON to belt app run falai/reve.

What output format does BiRefNet background removal return?

The skill returns a PNG file with a transparent background. This format preserves the cutout subject and works directly in design tools, storefronts, and marketing assets.

Do I need to install anything to use background removal with belt?

Yes, you need the inference.sh CLI (belt) installed and authenticated via belt login. Install instructions are linked in the skill documentation from the inference.sh skills repository.

Can I generate an image and remove its background in one workflow?

Yes, generate an image first with an app like falai/flux-dev-lora, then pass the resulting image URL to falai/reve or BiRefNet with a background removal prompt to get a transparent cutout.