← LeaderboardThe Skill Leaderboard

image-editing

Edit existing images with Bria, Reve and Phota via inference.sh CLI. Apps: Bria Eraser, Bria Gen Fill, Bria Expand, Bria Replace Background, Bria Increase Resolution, Bria Product Cutout, Bria Product Packshot, Bria Product Shadow, Bria Image Edit, Reve Edit, Reve Remix, Phota Edit, Phota Enhance. Capabilities: object removal, mask-based generative fill, outpainting, background replacement, instruction-based editing, multi-image remix, identity-preserving edits, photo enhancement, 2x/4x upscaling, product cutouts, packshots and shadows. Use for: photo retouching, e-commerce product images, rem

Published by inference-sh ↗inference-sh/skills

Try your first task

A source-grounded starter task has not been verified for this skill yet.

SKILL.md

name: image-editingdescription: "Edit existing images with Bria, Reve and Phota via inference.sh CLI. Apps: Bria Eraser, Bria Gen Fill, Bria Expand, Bria Replace Background, Bria Increase Resolution, Bria Product Cutout, Bria Product Packshot, Bria Product Shadow, Bria Image Edit, Reve Edit, Reve Remix, Phota Edit, Phota Enhance. Capabilities: object removal, mask-based generative fill, outpainting, background replacement, instruction-based editing, multi-image remix, identity-preserving edits, photo enhancement, 2x/4x upscaling, product cutouts, packshots and shadows. Use for: photo retouching, e-commerce product images, removing objects, changing backgrounds, extending images. Triggers: edit image, image editing, photo editing, remove object, erase object, inpainting, generative fill, outpainting, expand image, extend canvas, uncrop, replace background, change background, edit with text, instruction edit, bria, reve edit, reve remix, phota, enhance photo, increase resolution, product cutout, packshot, retouch"allowed-tools: Bash(belt *)

Install the belt CLI skill: npx skills add belt-sh/cli

Image Editing

Edit existing images via inference.sh CLI: remove objects, fill and extend, swap backgrounds, upscale, and prepare product shots.

Quick Start

Requires inference.sh CLI (belt). Install instructions

belt login

belt app run reve/edit --input '{
  "image": "https://your-photo.jpg",
  "edit_instruction": "Remove all people in the background"
}'

Apps by Task

TaskApp IDInputNotes
Remove objectsbria/eraseimage + maskWhite = erase, black = keep
Add / replace in a regionbria/gen-fillimage + mask + promptGenerates prompt content in the masked area
Edit with textbria/editimage + instructionNatural language edits
Edit with textreve/editimage + edit_instructionOptional upscale and transparent background
Edit with identityphota/edit1-10 images + promptKeeps trained subject identities, 1K or 4K output
Combine referencesreve/remix1-6 images + promptRefer to images as 0, 1, ... in the prompt
Background replacebria/replace-backgroundimage + prompt or ref_imageModes: fast, base, high_control
Background removereve/editimage + remove_background: trueTransparent output
Extend canvasbria/expandimage + aspect_ratio or canvas sizeOutpainting up to 5000x5000
Upscalebria/increase-resolutionimage2x or 4x, max 8192x8192
Enhance photophota/enhanceimageLighting, composition, color, sharpness
Product cutoutbria/product-cutoutproduct imageTransparent background, max 12MB
Product packshotbria/product-packshotproduct image or cutout2000x2000, hex or transparent background
Product shadowbria/product-shadowproduct cutoutRegular or floating shadow

Mask-based apps (bria/erase, bria/gen-fill) give exact control over the edited region. Instruction-based apps (bria/edit, reve/edit, phota/edit) need no mask.

Remove

Erase with a Mask

The mask must match the image aspect ratio.

belt app run bria/erase --input '{
  "image": "https://street-photo.jpg",
  "mask": "https://person-mask.png"
}'

Erase with an Instruction

belt app run reve/edit --input '{
  "image": "https://beach-photo.jpg",
  "edit_instruction": "Remove the trash can on the left"
}'

Add or Replace

Generative Fill

belt app run bria/gen-fill --input '{
  "image": "https://living-room.jpg",
  "mask": "https://empty-corner-mask.png",
  "prompt": "a tall green potted monstera plant",
  "negative_prompt": "blurry, distorted"
}'

Instruction Edit (Bria)

belt app run bria/edit --input '{
  "image": "https://portrait.jpg",
  "instruction": "change the shirt color to navy blue"
}'

Instruction Edit (Reve) with Upscale

test_time_scaling (1-5) spends more time for quality; values above 1 cost more.

belt app run reve/edit --input '{
  "image": "https://room.jpg",
  "edit_instruction": "Replace the sofa with a mid-century leather couch",
  "test_time_scaling": 2,
  "upscale_factor": 2
}'

Identity-Preserving Edit (Phota)

Reference trained profiles in the prompt with [[profile_id]].

belt app run phota/edit --input '{
  "prompt": "put [[my-profile]] in a tailored grey suit, studio lighting",
  "images": ["https://headshot.jpg"],
  "profile_ids": ["my-profile"],
  "resolution": "4K",
  "num_output_images": 2
}'

Remix Multiple Images

belt app run reve/remix --input '{
  "images": ["https://woman.jpg", "https://vintage-car.jpg"],
  "prompt": "The woman from 0 driving the car from 1 along a coastal road",
  "aspect_ratio": "16:9"
}'

Background

Replace Background from Text

belt app run bria/replace-background --input '{
  "image": "https://product-on-desk.jpg",
  "prompt": "a sunlit marble kitchen counter with soft morning light, blurred plants and a window in the background, warm natural tones",
  "mode": "high_control"
}'

Replace Background from a Reference Image

belt app run bria/replace-background --input '{
  "image": "https://portrait.jpg",
  "ref_image": "https://office-background.jpg"
}'

Transparent Background

belt app run reve/edit --input '{
  "image": "https://mascot.jpg",
  "edit_instruction": "Keep the mascot unchanged",
  "remove_background": true
}'

Extend the Canvas

belt app run bria/expand --input '{
  "image": "https://square-photo.jpg",
  "aspect_ratio": "16:9",
  "prompt": "continuation of the forest scene"
}'

Restore and Upscale

Upscale 4x

belt app run bria/increase-resolution --input '{
  "image": "https://small-image.jpg",
  "desired_increase": 4
}'

Enhance a Photo

belt app run phota/enhance --input '{
  "image": "https://dim-phone-photo.jpg",
  "output_format": "jpg"
}'

Product and E-commerce

Cutout

belt app run bria/product-cutout --input '{
  "image": "https://sneaker-photo.jpg"
}'

Packshot on White

belt app run bria/product-packshot --input '{
  "image": "https://sneaker-photo.jpg",
  "background_color": "#FFFFFF"
}'

Floating Shadow

belt app run bria/product-shadow --input '{
  "image": "https://sneaker-cutout.png",
  "shadow_type": "float",
  "shadow_intensity": 40
}'

Workflow: Raw Photo to Listing Image

# 1. Cut out the product
belt app run bria/product-cutout --input '{"image": "https://raw-product.jpg"}'

# 2. Add a natural drop shadow
belt app run bria/product-shadow --input '{
  "image": "<cutout-url-from-step-1>",
  "shadow_type": "regular"
}'

# 3. Upscale for zoom views
belt app run bria/increase-resolution --input '{
  "image": "<url-from-step-2>",
  "desired_increase": 2
}'

Pricing

AppPrice
bria/erase$0.02/image
bria/expand$0.02/image
bria/increase-resolution$0.02/image
bria/product-cutout$0.018/image
bria/product-shadow$0.018/image
bria/gen-fill$0.03/image
bria/replace-background$0.03/image
bria/product-packshot$0.03/image
bria/edit$0.03/image
phota/edit$0.09/image standard, $0.18/image 4K
phota/enhance$0.13/image

Check any app with belt app pricing <app-id>.

Tips

  • Get the full input schema for any app: belt app sample <app-id>
  • bria/replace-background prompts work best in English at 50-110 words.
  • bria/expand without aspect_ratio needs canvas_size, original_image_size and original_image_location.
  • Pass a seed (seed on Bria apps) to reproduce a result.

Related Skills

# Full platform skill (all apps)
npx skills add inference-sh/skills@infsh-cli

# Image generation
npx skills add inference-sh/skills@ai-image-generation

# Background removal
npx skills add inference-sh/skills@background-removal

# Upscaling
npx skills add inference-sh/skills@image-upscaling

# Product photography
npx skills add inference-sh/skills@ai-product-photography
npx skills add inference-sh/skills@product-photography

Browse all image apps: belt app list --category image

Documentation