tilltheend/unwrap-clothes
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
No description
11 commits 1 branch 0 tags 189 KiB
Find a file
Schmidt Till (CSS TO PME DSI PAO MUC) 6650298c3d refactor: address review feedback on multipart field compatibility and form serialization
2026-09-10 23:32:40 +02:00
.agents/skills chore: add agent skills, docs, and project context 2026-09-10 23:12:00 +02:00
docs docs(map): record resolution of Implement FastAPI backend and /api/clean endpoint in app.py 2026-09-10 23:26:55 +02:00
static refactor: address review feedback on multipart field compatibility and form serialization 2026-09-10 23:32:40 +02:00
tests feat: build single-page web UI in static/index.html (#4) 2026-09-10 23:28:59 +02:00
.gitignore chore: add agent skills, docs, and project context 2026-09-10 23:12:00 +02:00
AGENTS.md chore: add agent skills, docs, and project context 2026-09-10 23:12:00 +02:00
app.py refactor: address review feedback on multipart field compatibility and form serialization 2026-09-10 23:32:40 +02:00
cleaner.py refactor: address code review feedback on public image info, API key resolution, and serialization 2026-09-10 23:26:36 +02:00
CONTEXT.md chore: add agent skills, docs, and project context 2026-09-10 23:12:00 +02:00
gemini-only.yml chore: add agent skills, docs, and project context 2026-09-10 23:12:00 +02:00
mise.toml chore: add agent skills, docs, and project context 2026-09-10 23:12:00 +02:00
pytest.ini chore: add pytest.ini with pythonpath for direct test execution 2026-09-10 23:15:54 +02:00
README.md Initial commit: garment photo unwrinkle and background cleanup tool 2026-09-10 15:26:38 +02:00
requirements.txt feat: implement FastAPI backend and /api/clean endpoint (#3) 2026-09-10 23:24:36 +02:00
skills-lock.json chore: add agent skills, docs, and project context 2026-09-10 23:12:00 +02:00
unwrap_clothes.py refactor: address code review feedback on image info extraction and rescaling 2026-09-10 23:20:47 +02:00

README.md

Garment photo cleaner

CLI tool that cleans up second-hand clothing photos for marketplace listings (Vinted, eBay, Depop). It removes fabric wrinkles, puts the garment on a studio backdrop (warm beige or bright white), and preserves shapes, colors, prints, buttons, and defects.

Uses OpenRouter's Image API with meta/muse-image.

Requirements

  • Python 3.9 or higher
  • OpenRouter API key with image generation credits

Install dependencies:

pip install -r requirements.txt

Set your API key:

export OPENROUTER_API_KEY="sk-or-v1-..."

Usage

Basic run with default soft beige background:

python3 unwrap_clothes.py shirt.jpg

Output saves next to the source photo as shirt_clean.jpg.

Options

Select a white background:

python3 unwrap_clothes.py shirt.jpg --bg white

Custom output path:

python3 unwrap_clothes.py shirt.jpg -o /path/to/listing_photo.jpg

Custom background description:

python3 unwrap_clothes.py shirt.jpg --bg "Soft cool grey studio background, even lighting"

Keep native model resolution (skip automatic upscale to input dimensions):

python3 unwrap_clothes.py shirt.jpg --no-restore-res

Override the model:

python3 unwrap_clothes.py shirt.jpg --model meta/muse-image

Why meta/muse-image is the default

Other image models on OpenRouter run into policy or cost issues on second-hand clothing photos:

  1. openai/gpt-5-image-mini: OpenAI's input image filter rejects photos with licensed character artwork (such as Winnie the Pooh, Disney, or cartoon prints), returning HTTP 400 safety errors regardless of prompt phrasing.
  2. qwen/qwen-image-3: Alibaba moderation frequently blocks full garment photos, and Qwen 3 Pro costs roughly $0.08 per image.
  3. meta/muse-image: Accepts photos of branded and character clothing, costs $0.01 per image, and finishes in 15 to 25 seconds.

Resolution handling

meta/muse-image caps output at roughly 1.3 to 1.8 MP (for example, 1376x1824) regardless of the size parameters passed to the API.

To avoid downsized uploads on marketplaces that expect high-resolution smartphone photos (such as 3024x4032), the script automatically resizes the generated image back to the source file's exact dimensions using Lanczos interpolation. This preserves the aspect ratio and frame size of the original photo.

Known limitations

  • Garments that are physically folded or bunched over themselves in the photo remain bunched. Generative editing smooths surface wrinkles, but will not unfold fabric that covers other parts of the garment.
  • Small text on labels and complex repeating patterns can drift slightly from the original during generation.