# REST API Documentation
This REST API provides HTTP access to the Content Credentials verification functionality, making it accessible to ChatGPT custom GPTs and other web services.
## Quick Start
### 1. Install Dependencies
```bash
npm install
```
### 2. Build the Project
```bash
npm run build
```
### 3. Start the REST API Server
```bash
npm run start:api
```
Or for development with auto-reload:
```bash
npm run dev:api
```
The server will start on `http://localhost:3000` by default.
### 4. Test the API
```bash
# Health check
curl http://localhost:3000/health
# Verify an image
curl "http://localhost:3000/verify-url?url=https://example.com/image.jpg"
```
## Endpoints
### `GET /`
API documentation and available endpoints.
**Response:**
```json
{
"name": "Content Credentials API",
"version": "1.0.0",
"description": "REST API for reading C2PA Content Credentials",
"endpoints": { ... }
}
```
### `GET /health`
Health check endpoint.
**Response:**
```json
{
"status": "ok",
"service": "content-credentials-api",
"version": "1.0.0",
"timestamp": "2025-12-11T10:00:00.000Z"
}
```
### `GET /.well-known/openapi.json`
OpenAPI 3.0 schema for ChatGPT integration.
**Response:** Full OpenAPI specification
### `GET /verify-url`
Verify Content Credentials from a URL (query parameter).
**Parameters:**
- `url` (query, required): The URL of the image or video to verify
**Example:**
```bash
curl "http://localhost:3000/verify-url?url=https://example.com/image.jpg"
```
**Response:**
```json
{
"success": true,
"hasCredentials": true,
"manifestData": {
"whoThisComesFrom": { ... },
"aboutThisContent": { ... },
"aboutTheseCredentials": { ... },
"validationInfo": { ... }
}
}
```
### `POST /verify-url`
Verify Content Credentials from a URL (JSON body).
**Body:**
```json
{
"url": "https://example.com/image.jpg"
}
```
**Example:**
```bash
curl -X POST http://localhost:3000/verify-url \
-H "Content-Type: application/json" \
-d '{"url":"https://example.com/image.jpg"}'
```
**Response:** Same as GET `/verify-url`
## ChatGPT Integration
### Using with ChatGPT Custom GPTs
1. **Start the server:**
```bash
npm run start:api
```
2. **In ChatGPT GPT Builder:**
- Go to "Configure" → "Actions"
- Click "Create new action"
- Import schema from: `http://localhost:3000/.well-known/openapi.json`
- **Or** copy/paste the OpenAPI schema directly
3. **Important:** ChatGPT needs to be able to reach your local server:
- Use a tunneling service like **ngrok** to expose your local server
- Or deploy to a public server (Render, Railway, Heroku, etc.)
### Using ngrok (Recommended for Testing)
```bash
# Install ngrok
brew install ngrok
# Start your API server
npm run start:api
# In another terminal, create a tunnel
ngrok http 3000
```
ngrok will give you a public URL like `https://abc123.ngrok.io`. Use this URL in ChatGPT instead of `http://localhost:3000`.
## Response Format
### Success Response (with credentials)
```json
{
"success": true,
"hasCredentials": true,
"manifestData": {
"whoThisComesFrom": {
"linkedInIdentity": {
"name": "John Doe",
"profileUrl": "https://linkedin.com/in/johndoe",
"verified": true
},
"otherIdentities": [
{
"name": "Adobe Firefly",
"socialAccounts": ["@adobe"]
}
]
},
"aboutThisContent": {
"actions": [
{
"action": "c2pa.created",
"softwareAgent": "Adobe Firefly",
"when": "2024-12-11T10:00:00Z"
}
],
"genAIInfo": {
"generative": true,
"training": false,
"model": "Adobe Firefly Image 3"
}
},
"aboutTheseCredentials": {
"claimSigner": "Adobe Inc.",
"timestamp": "2024-12-11T10:00:00Z"
},
"validationInfo": {
"certificate": {
"issuer": "ContentAuthenticity",
"serialNumber": "abc123"
},
"trustInfo": {
"isValid": true
}
}
}
}
```
### Success Response (no credentials)
```json
{
"success": true,
"hasCredentials": false
}
```
### Error Response
```json
{
"success": false,
"hasCredentials": false,
"error": "Invalid URL format"
}
```
## Environment Variables
- `PORT` - Server port (default: 3000)
- `LOG_LEVEL` - Logging level: `debug`, `info`, `warn`, `error` (default: `info`)
## Deployment
### Deploy to Render.com
1. Create a `render.yaml`:
```yaml
services:
- type: web
name: content-credentials-api
env: node
buildCommand: npm install && npm run build
startCommand: npm run start:api
```
2. Connect your GitHub repo to Render
3. Deploy!
### Deploy to Railway
1. Install Railway CLI:
```bash
npm install -g @railway/cli
```
2. Deploy:
```bash
railway login
railway init
railway up
```
### Deploy to Heroku
```bash
heroku create your-app-name
git push heroku main
```
## Security Considerations
- This API only supports URL-based verification (no file uploads)
- CORS is enabled for all origins (configure as needed)
- Input validation on all endpoints
- No authentication by default (add as needed for production)
## Limitations
- **Local files:** This API cannot verify local files (by design, for security)
- **URL-only:** Images must be accessible via HTTP/HTTPS URLs
- **Dependencies:** Requires c2patool and TrustMark to be installed
## Support
For issues or questions:
- GitHub: https://github.com/noga7/mcp-content-credentials
- Issues: https://github.com/noga7/mcp-content-credentials/issues
