io.github.jaupin94/lassocut-mcp
Remove image backgrounds from AI assistants with a LassoCut account (files, folders, URLs).
Install
Add this to your client's MCP configuration (Claude Desktop: claude_desktop_config.json; Cursor: .cursor/mcp.json; VS Code and others use the same "mcpServers" shape).
lassocut-mcpclaude mcp add lassocut-mcp -e 'LASSOCUT_API_KEY=<your lassocut_api_key>' -- npx -y lassocut-mcp@1.1.0
{
"mcpServers": {
"lassocut-mcp": {
"command": "npx",
"args": [
"-y",
"lassocut-mcp@1.1.0"
],
"env": {
"LASSOCUT_API_KEY": "<your lassocut_api_key>"
}
}
}
}
Environment variables
LASSOCUT_API_KEYsecret Optional LassoCut API key (otherwise use the sign_in tool)
Documentation
From the project's README on GitHub, refreshed weekly. Links open on GitHub.
LassoCut MCP
Remove image backgrounds from your AI assistant with a LassoCut account. The API is compatible with the remove.bg API.
What it does
Ask Claude (or any MCP-compatible assistant) to cut out product photos, portraits, whole folders, or images from a URL. Results are saved next to the originals as <name>-no-bg.<ext> and existing files are never overwritten — a name already taken gets -2, -3, and so on instead.
By default, every image is processed as a low-cost preview (0.25 credit, 50 free a month, up to 10 a day). Full-size output costs 1 credit per image and is only used when you explicitly ask for it; above 10 full-size images, the assistant must show you the cost and get your OK first.
Install
Claude Desktop: download lassocut.mcpb and double-click it.
Claude Code:
claude mcp add lassocut -- npx -y lassocut-mcp
On Windows: claude mcp add lassocut -- cmd /c npx -y lassocut-mcp
Cursor / Windsurf (or any other MCP client that reads a JSON config):
{ "mcpServers": { "lassocut": { "command": "npx", "args": ["-y", "lassocut-mcp"] } } }
Requires Node.js 20 or later.
Sign in
You need a LassoCut account and API key. Three ways to connect:
- Ask the assistant to sign in. It calls the
sign_intool, which opens the LassoCut approval page in your browser and returns the approval link and a short code right away. Approve the connection in the browser, then ask the assistant to check your credits (get_credits) — the key is saved automatically in the background (the approval window stays open for a few minutes). - Already signed in with the
lassocutcommand-line tool? This connector reads the same saved key fromlassocut login, so no extra step is needed. - Set the
LASSOCUT_API_KEYenvironment variable (or theapi_keyfield in Claude Desktop's extension settings) to a key from your LassoCut account page. This always takes precedence over a signed-in key.
Tools
remove_background
Remove the background of one or more images.
| Parameter | Description |
|---|---|
images | A file path, a folder path (its images, not subfolders), an http(s) URL, or a list of any of these. Up to 50 per call. |
size | preview (default) or full. |
background | transparent (default), a hex color such as #ffffff, or a color name. |
format | png (default), jpg, or webp. |
crop | Crop the result to the subject. |
preset | none (default), amazon, ebay, etsy, shopify, zalando or instagram. A preset sets the background, format and crop itself (those arguments are ignored), saves a .jpg, and reports any file that is not compliant. Use size full for marketplace uploads; a preview is enlarged and looks soft. Compliant means compliant with the image rules LassoCut checks. |
shadow | none (default) or cast. Not available with the zalando preset. |
light_angle | Shadow light angle, -90 to 90 (positive: light from the left). Default 45. |
shadow_length | short, medium or long. |
shadow_strength | Shadow strength, 0 to 100 (soft 35, normal 55, strong 75). |
output_dir | Where to save results. Must be an absolute path (~ is expanded). If omitted: next to each original file, or your Downloads folder (falling back to your home folder) for images fetched from a URL. |
confirm_cost | Set to true only after the user has confirmed the cost of a full-size request over 10 images. |
Files larger than 22 MB are skipped. Up to 3 small thumbnails are returned inline with the result.
get_credits
Shows your current credit balance and the free previews available now (50 a month, up to 10 a day).
sign_in
Connects a LassoCut account: opens the approval page in your browser and returns immediately with the link and code so you can approve at your own pace. Pass force: true to switch to a different account.
Credits and safety
- Previews are the default everywhere; full-size output is never used unless asked for.
- Requesting full size for more than 10 images at once requires an explicit confirmation, with the cost shown up front.
- If your balance is too low for a full-size request, it is refused before any image is sent.
- Files are only ever sent to the LassoCut API — see Privacy below.
Privacy
Images are sent only to the LassoCut API to remove their background; they are not sent anywhere else. See the LassoCut privacy policy for details.
remove.bg is a trademark of Canva Austria GmbH. LassoCut is an independent product of JAUPIN Design LLC and is not affiliated with, endorsed by, or sponsored by Canva or remove.bg.
License
MIT — see LICENSE.
Frequently asked questions
What is io.github.jaupin94/lassocut-mcp?
Remove image backgrounds from AI assistants with a LassoCut account (files, folders, URLs).
How do I install io.github.jaupin94/lassocut-mcp?
It runs on your machine. In Claude Code: claude mcp add lassocut-mcp -e 'LASSOCUT_API_KEY=<your lassocut_api_key>' -- npx -y lassocut-mcp@1.1.0. For other clients, paste the JSON configuration from this page into their MCP settings.
Does io.github.jaupin94/lassocut-mcp need an API key?
Yes. It expects LASSOCUT_API_KEY, which you set in its configuration.
Is io.github.jaupin94/lassocut-mcp open source?
Yes. Its code is on GitHub under the MIT license.
Who maintains io.github.jaupin94/lassocut-mcp?
It is published by jaupin94. AI.tn lists it but does not maintain or endorse it.
From the official MCP Registry. Listing is not an endorsement: review a server's code and permissions before giving it access to your data.