- Home
- Skills
- E-commerce
- Local Background Remover
More screenshots
Works with the AI Agents you already use
Local Background Remover
Remove image backgrounds on your own computer: unlimited, private, no API keys, no per-image fees.
$9.99
Local Background Remover
Example session with this skill installed
Remove backgrounds from all images in ~/Shop/new-arrivals and put them on white with a floor shadow, as JPEG.
- Read your context and instructions
- Compiled the local background remover
Done: 48 processed, 0 failed (standard model, NVIDIA GPU). Saved to ~/Shop/new-arrivals/Background-Removed/ as *-no-bg.jpg on white with a soft floor shadow. Report: bgremover-report.json.
Connects securely to your tools. The creator never sees your data.
About this skill
Stop paying per image and uploading your photos to someone else's server. This skill gives your AI agent a local background-removal engine. Your agent installs it with you (it asks first), then cuts out as many images as you want, 100% offline.
What it does
- Transparent PNG cutouts with a real alpha channel, at your photo's exact size
- Clean edges: built-in edge colour decontamination removes the halo of the old background around hair and soft edges
- Batch whole folders (subfolders too). It resumes where it stopped, one corrupt file never kills the run, and you get a JSON report
- Replace backgrounds with white, any colour, or another photo: product-on-white JPEGs in one command
- Floor and drop shadows for product shots
- Two MIT-licensed AI models (BiRefNet-lite and BiRefNet): commercial use of your results is OK
- CPU or NVIDIA GPU. It detects your hardware, picks the right CUDA runtime for your driver automatically, and needs no CUDA Toolkit install. In testing, a 1080p photo took ~1.3 s on an RTX 4060 Ti and ~3-6 s on CPU.
Built to be trusted
- Shows you a plan (what's downloaded, how big, where it goes) and waits for your yes
- Installs into one folder: no admin rights, your system Python untouched, one-command uninstall
- Models are downloaded from the original publishers at pinned versions and SHA-256 verified
- Never overwrites or deletes your originals
- Checks every output it writes (size, transparency, empty or failed cutouts), and tells you honestly when a photo is hard
Try it with
- "Install Local Background Remover."
- "Remove the background from IMG_2041.jpg."
- "Put every photo in Products/ on a white background with a soft shadow, as JPEG."
- "Use the best quality model on this portrait. The hair is tricky."
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for E-commerce.
Very fine hair, glass, smoke and subjects that blend into the background may not be perfect. The Advanced model helps, but no model is flawless. HEIC/RAW files need converting to JPG/PNG first. macOS runs on CPU (no CUDA on Mac). First setup needs internet and Python 3.10+.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed today
- Accounts
Hugging Face
- Software
Python 3.10+
Windows, macOSorLinux- Optional
- NVIDIA GPU (driver 525+)
One-time setup downloads ~60 MB of Python packages (CPU) or ~1.3 GB (NVIDIA GPU), plus a 224 MB model. After that, everything runs offline.
Permissions
Allowed Hosts
File Scopes
Network is used only during setup steps the user approves (Python packages from PyPI, pinned model weights from Hugging Face). Image processing never touches the network. No API keys, no telemetry.