Content Credentials

You can use the MCP Content Credentials Server to read C2PA content credentials from images and videos, either from embedded manifests or trust-based watermarks. It combines fast embedding-detection with a fallback watermark check and exposes a REST API for integration with ChatGPT and web apps, making it practical to verify content provenance at scale.

How to use

Install and run the server locally, then connect your MCP client or REST API consumer to perform checks on images, videos, or URLs. You can check a specific file, browse directories, or verify a URL. The server will first attempt to read embedded C2PA manifests for a fast result, then fall back to detecting TrustMark watermarks if needed. When credentials are found, you receive structured, human-readable data about the source, issuer, and content provenance.

How to install

Prerequisites you need before installation:

• Node.js v18 or newer

• Python 3.8.5 or newer (for TrustMark watermarks)

Follow these steps to set up the MCP Content Credentials Server locally:

# 1. Clone
git clone https://github.com/noga7/mcp-content-credentials.git
cd mcp-content-credentials

# 2. Install dependencies (automatic: installs c2patool + TrustMark)
npm install

# 3. Build
npm run build

Additional configuration and starting the REST API

Configure the local Claude Desktop integration by adding the MCP server mapping to your Claude configuration. Create or update the following JSON snippet to point to the built index file.

{
  "mcpServers": {
    "content-credentials": {
      "command": "node",
      "args": ["/absolute/path/to/mcp-content-credentials/build/index.js"]
    }
  }
}

Restart Claude Desktop after saving the config. This loads the MCP server so you can issue queries like:

• Check a specific image file
• Browse directories for credentials
• Verify a URL directly via the MCP API

Starting the REST API for ChatGPT and web apps

If you want to access the server via HTTP for ChatGPT or your own web app, start the REST API. The API runs locally on port 3000 by default.

npm run start:api

# Quick test
curl "http://localhost:3000/verify-url?url=https://example.com/image.jpg"

Available tools

Embedded C2PA Detection

Reads manifests from file metadata to extract content credential data quickly.

TrustMark Watermark Detection

Detects credentials embedded in image pixels as a fallback when no embedded manifest is found.

REST API

Provides HTTP endpoints for ChatGPT and web app integration to verify credentials and retrieve structured results.

Filesystem Access

Allows direct browsing of user directories to locate potential credential-bearing media.

Automatic Installation

Automates dependency installation so you can get up and running with minimal setup.

Structured Output

Produces human-readable parsed credential data for easy consumption by clients.