ImgVX AI MCP Server

Name: ImgVX AI
Author: imgvx

Official Model Context Protocol server for ImgVX AI, free image tools that run in the browser. Connect it to Claude, ChatGPT, Cursor, VS Code, Windsurf, Zapier, Make, n8n or any other MCP client to:

Find the right image tool for a job such as removing a background, upscaling, erasing an object, enhancing or colorizing a photo, compressing to a file size, resizing, cropping or converting HEIC, PNG, JPG, WebP or PDF.
Get a link in the user's language: the most-used tools have pages in 40 languages.
List every tool (87, in 8 categories) and check pricing (most tools are free; HD downloads use credits).

It is a hosted server: there is nothing to install, no sign-in and no API key. The tools never touch images. They return a link, and the user edits the image in their own browser, where it is never uploaded.


URL	`https://imgvx.com/mcp`
Transport	Streamable HTTP
Authentication	None
Tools	3, all read-only
Registry name	`com.imgvx/imgvx-ai`
Docs	https://imgvx.com/mcp-server/

Setup

Claude

In Claude (web or desktop), open Settings → Connectors → Add custom connector, name it ImgVX AI and paste https://imgvx.com/mcp.

Claude Code:

bash

claude mcp add --transport http imgvx-ai https://imgvx.com/mcp

ChatGPT

With developer mode on (Settings → Apps & Connectors → Advanced), choose Create, name it ImgVX AI, paste https://imgvx.com/mcp and pick No authentication.

Cursor

~/.cursor/mcp.json (or .cursor/mcp.json in a project):

json

{
  "mcpServers": {
    "imgvx-ai": { "url": "https://imgvx.com/mcp" }
  }
}

VS Code (GitHub Copilot)

.vscode/mcp.json:

json

{
  "servers": {
    "imgvx-ai": { "type": "http", "url": "https://imgvx.com/mcp" }
  }
}

Windsurf

~/.codeium/windsurf/mcp_config.json:

json

{
  "mcpServers": {
    "imgvx-ai": { "serverUrl": "https://imgvx.com/mcp" }
  }
}

Zapier, Make and n8n

Zapier: add an MCP Client step, connect with the server URL, transport Streamable HTTP, no authentication.
Make: add the MCP Client module, create a connection with the server URL and no authentication.
n8n: give an AI Agent node the MCP Client Tool, set the endpoint to the server URL, transport HTTP Streamable, authentication None.

Tools

Tool	Input	Returns
`find_image_tool`	`task` (what to do to the image, in a few words), optional `language` (one of 40 language codes)	Up to five matching tools with what each does, its price and a link, in the user's language where a page exists
`list_image_tools`	optional `category` (`ai`, `compress-resize`, `adjust`, `effects`, `edit`, `convert`, `inspect`, `create`)	Every tool in that category with a short description and a link
`get_pricing`	none	What is free and what Pro and the credit packs cost

Example prompts

"Remove the background from my product photo, which free tool should I use?"
"I need to compress a passport photo to under 100 KB."
"Make this old photo 4x bigger without it going blurry."
"¿Cómo quito un objeto de una foto gratis?"

Try it with curl

bash

curl -s https://imgvx.com/mcp \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"find_image_tool","arguments":{"task":"remove background","language":"es"}}}'

Privacy and support

The server receives only the tool inputs your app sends. It does not store them or link them to a person, and sets no cookies. Images are never sent to it: every ImgVX AI tool runs in the user's browser. See the privacy policy.

Questions or problems: support@imgvx.com or imgvx.com/contact.

About this repository

This repository holds the public listing (server.json for the MCP Registry) and setup docs for the hosted ImgVX AI MCP server, under the MIT license. The server itself runs at imgvx.com.