🔌 MCP Server Documentation

Connect AI Watermark Remover to your AI coding assistant via the Model Context Protocol (MCP). Detect and remove AI watermarks directly from Claude, Cursor, Windsurf, and more.

● Online SSE Transport 4 Tools Free

Server Information

Property	Value
Server Name	`ai-watermark-remover`
SSE Endpoint	`https://www.aiwatermarksremover.com/sse`
Transport	Server-Sent Events (SSE)
Authentication	None required (free & open)
Rate Limit	30 requests/minute per IP

Connection Methods

🖥️ Claude Desktop App

Edit your claude_desktop_config.json file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

claude_desktop_config.json

{
  "mcpServers": {
    "ai-watermark-remover": {
      "url": "https://www.aiwatermarksremover.com/sse"
    }
  }
}

⌨️ Claude Code (CLI)

Add to your project or global settings:

.claude/settings.json

{
  "mcpServers": {
    "ai-watermark-remover": {
      "type": "sse",
      "url": "https://www.aiwatermarksremover.com/sse"
    }
  }
}

Or add via CLI command:

claude mcp add ai-watermark-remover --transport sse https://www.aiwatermarksremover.com/sse

🌐 Claude.ai Web / Desktop MCP Panel

Click "+ 添加服务器 / + Add Server" and fill in:

Option A: Using npx (requires Node.js)

Field	Value
Name (名称)	`ai-watermark-remover`
Command (命令)	`npx`
Args (参数)	`-y mcp-remote https://www.aiwatermarksremover.com/sse`
ENV (环境变量)	leave empty

⚠️ "Error: spawn npx ENOENT"? This means Node.js is not installed. Install it from nodejs.org, then restart Claude and try again. Or use Option B below.

Option B: Using uvx (requires uv / Python)

If you don't have Node.js but have uv installed:

Field	Value
Name (名称)	`ai-watermark-remover`
Command (命令)	`uvx`
Args (参数)	`mcp-remote https://www.aiwatermarksremover.com/sse`
ENV (环境变量)	leave empty

Option C: Using Python directly

If you have Python 3.10+ installed:

# Install mcp-remote first
pip install mcp-remote

Field	Value
Name (名称)	`ai-watermark-remover`
Command (命令)	`python`
Args (参数)	`-m mcp_remote https://www.aiwatermarksremover.com/sse`
ENV (环境变量)	leave empty

Tip: On macOS/Linux, you may need to use python3 instead of python.

📝 Cursor IDE

Go to Settings → MCP Servers, add a new server:

~/.cursor/mcp.json

{
  "mcpServers": {
    "ai-watermark-remover": {
      "url": "https://www.aiwatermarksremover.com/sse"
    }
  }
}

Or use the command bridge method:

{
  "mcpServers": {
    "ai-watermark-remover": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://www.aiwatermarksremover.com/sse"]
    }
  }
}

🏄 Windsurf (Codeium)

Add to your Windsurf MCP configuration:

~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "ai-watermark-remover": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://www.aiwatermarksremover.com/sse"]
    }
  }
}

🔗 Any MCP-Compatible Client

Connect using the SSE endpoint URL:

SSE Endpoint: https://www.aiwatermarksremover.com/sse
Messages:     https://www.aiwatermarksremover.com/messages/

For stdio-only clients, use the npx bridge:

npx -y mcp-remote https://www.aiwatermarksremover.com/sse

Available Tools

detect_watermark

Detect hidden AI watermarks in text. Scans for invisible Unicode characters (zero-width spaces, joiners, bidirectional marks) embedded by ChatGPT, Claude, Gemini, and other AI models.

text REQUIRED string — The text to scan for hidden watermarks

remove_watermark

Remove hidden AI watermarks from text. Strips invisible Unicode characters and returns clean text with a removal report.

text REQUIRED string — The text to remove watermarks from

detect_file_watermark

Detect watermarks and metadata in a file. Scans for C2PA provenance manifests, EXIF/XMP metadata. Supports PNG, JPEG, WebP, SVG, PDF, DOCX, ODT.

file_base64 REQUIRED string — Base64-encoded file content

filename REQUIRED string — Original filename with extension (e.g. 'photo.png')

remove_file_watermark

Remove watermarks and metadata from a file. Strips C2PA manifests, EXIF/XMP tags. Returns cleaned file as base64.

file_base64 REQUIRED string — Base64-encoded file content

filename REQUIRED string — Original filename with extension

Usage Examples

In Claude Desktop / Claude Code

Once connected, simply ask Claude to use the tools:

# Detect watermarks
"Check this text for hidden AI watermarks: [paste your text]"

# Remove watermarks
"Remove any hidden watermarks from this text: [paste your text]"

# File watermark detection
"Detect watermarks in this PDF file"

Programmatic (Python)

Python MCP Client

from mcp import ClientSession
from mcp.client.sse import sse_client

async def main():
    async with sse_client("https://www.aiwatermarksremover.com/sse") as (read, write):
        async with ClientSession(read, write) as session:
            await session.initialize()

            # List tools
            tools = await session.list_tools()
            print(tools)

            # Detect watermarks
            result = await session.call_tool(
                "detect_watermark",
                {"text": "Your AI-generated text here"}
            )
            print(result)

FAQ

Q: Do I need an API key?
No. The MCP server is free and open. No authentication required.

Q: What's the rate limit?
30 requests per minute per IP address. Sufficient for normal usage.

Q: Is my data stored?
No. Text and files are processed in memory and never stored permanently.

Q: The npx command fails — what do I do?
Ensure Node.js (v18+) is installed. Run npm install -g mcp-remote first, then use mcp-remote https://www.aiwatermarksremover.com/sse as the command.