CUA Website Scan - KYC API Documentation

AI-powered website analysis using Computer Use Agent (CUA) technology. The agent autonomously navigates and analyzes websites to identify potential fraud, scams, compliance issues, and suspicious content.

Overview

The CUA Scan API uses Claude’s Computer Use capabilities to:

Navigate websites like a human user
Analyze content for fraud indicators, scams, and compliance issues
Capture screenshots at each step for audit trails
Generate detailed reports with findings and recommendations

CUA scans are asynchronous. Submit a job, then poll for status and results.

Endpoints

Endpoint	Method	Description
`/kyc/cua-scan`	POST	Submit a new scan job
`/kyc/cua-scan`	GET	List jobs with pagination
`/kyc/cua-scan/{jobId}`	GET	Get job status and results

Authentication

Endpoint	Required Permission
Submit Job	`cua:create`
List Jobs	`cua:read`
Get Status	`cua:read`

Submit Scan Job

Create a new CUA scan job to analyze a website.

POST /kyc/cua-scan

Request Parameters

Effort Levels

Level	Steps	Use Case
`low`	20	Quick surface-level scan
`medium`	50	Standard comprehensive scan (default)
`high`	80	Deep analysis for complex sites

Request Example

curl -X POST https://stg.kyc.legaltalent.ai/kyc/cua-scan \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://suspicious-shop.com",
    "effort": "medium"
  }'

Response

{
  "job_id": "550e8400-e29b-41d4-a716-446655440000",
  "status": "PENDING",
  "target_url": "https://suspicious-shop.com",
  "effort": "medium",
  "created_at": "2024-12-15T10:30:00.000000+00:00",
  "message": "Job submitted successfully"
}

Response Fields

Field	Type	Description
`job_id`	string	Unique job identifier (UUID)
`status`	string	Initial status: `PENDING`
`target_url`	string	URL being scanned
`effort`	string	Effort level selected
`created_at`	string	ISO 8601 timestamp
`message`	string	Confirmation message

List Jobs

Get a paginated list of scan jobs for your tenant.

GET /kyc/cua-scan

Query Parameters

Parameter	Type	Default	Description
`limit`	integer	20	Number of items per page (max 100)
`cursor`	string	-	Pagination cursor from previous response
`status`	string	-	Filter by status: `PENDING`, `PROCESSING`, `COMPLETED`, `FAILED`
`sort`	string	`desc`	Sort order: `desc` (newest first) or `asc` (oldest first)

Request Example

# List recent jobs
curl -X GET "https://stg.kyc.legaltalent.ai/kyc/cua-scan?limit=10&sort=desc" \
  -H "Authorization: Bearer YOUR_TOKEN"

# Filter by status
curl -X GET "https://stg.kyc.legaltalent.ai/kyc/cua-scan?status=COMPLETED&limit=20" \
  -H "Authorization: Bearer YOUR_TOKEN"

# Paginate through results
curl -X GET "https://stg.kyc.legaltalent.ai/kyc/cua-scan?cursor=eyJqb2JfaWQiOiAiLi4uIn0=" \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

{
  "items": [
    {
      "job_id": "550e8400-e29b-41d4-a716-446655440000",
      "target_url": "https://suspicious-shop.com",
      "status": "COMPLETED",
      "effort": "medium",
      "created_at": "2024-12-15T10:30:00.000000+00:00",
      "updated_at": "2024-12-15T10:35:00.000000+00:00"
    },
    {
      "job_id": "660e8400-e29b-41d4-a716-446655440001",
      "target_url": "https://another-site.com",
      "status": "PROCESSING",
      "effort": "high",
      "created_at": "2024-12-15T10:32:00.000000+00:00",
      "updated_at": "2024-12-15T10:33:00.000000+00:00"
    }
  ],
  "count": 2,
  "cursor": "eyJqb2JfaWQiOiAiNjYwZTg0MDAuLi4ifQ==",
  "has_more": true
}

Response Fields

Field	Type	Description
`items`	array	List of job summaries
`count`	integer	Number of items in this response
`cursor`	string	Pagination cursor for next page (null if no more)
`has_more`	boolean	Whether more pages exist

Get Job Status

Get detailed status, progress, and results of a scan job.

GET /kyc/cua-scan/{jobId}

Path Parameters

Parameter	Type	Description
`jobId`	string	Job ID (UUID)

Query Parameters

Parameter	Type	Default	Description
`from_step`	integer	0	Return only steps with `sequence_id >= from_step` (for polling)

Request Example

# Get full status
curl -X GET "https://stg.kyc.legaltalent.ai/kyc/cua-scan/550e8400-e29b-41d4-a716-446655440000" \
  -H "Authorization: Bearer YOUR_TOKEN"

# Poll for new steps (efficient for real-time updates)
curl -X GET "https://stg.kyc.legaltalent.ai/kyc/cua-scan/550e8400-e29b-41d4-a716-446655440000?from_step=15" \
  -H "Authorization: Bearer YOUR_TOKEN"

Response - Processing

{
  "job_id": "550e8400-e29b-41d4-a716-446655440000",
  "target_url": "https://suspicious-shop.com",
  "status": "PROCESSING",
  "effort": "medium",
  "created_at": "2024-12-15T10:30:00.000000+00:00",
  "updated_at": "2024-12-15T10:31:30.000000+00:00",
  "current_step": 5,
  "total_sequences": 12,
  "last_sequence_id": 12,
  "from_sequence": 0,
  "steps": [
    {
      "sequence_id": 1,
      "agent_step": 1,
      "type": "text",
      "content": "I'll analyze this website for potential fraud indicators..."
    },
    {
      "sequence_id": 2,
      "agent_step": 1,
      "type": "tool_use",
      "tool": "computer",
      "action": "screenshot"
    },
    {
      "sequence_id": 3,
      "agent_step": 1,
      "type": "tool_result",
      "success": true,
      "screenshot_url": "https://bucket.s3.amazonaws.com/screenshots/step1.png?..."
    },
    {
      "sequence_id": 4,
      "agent_step": 2,
      "type": "text",
      "content": "The homepage shows several red flags: no company address..."
    }
  ]
}

Response - Completed

{
  "job_id": "550e8400-e29b-41d4-a716-446655440000",
  "target_url": "https://suspicious-shop.com",
  "status": "COMPLETED",
  "effort": "medium",
  "created_at": "2024-12-15T10:30:00.000000+00:00",
  "updated_at": "2024-12-15T10:35:00.000000+00:00",
  "current_step": 50,
  "total_sequences": 145,
  "last_sequence_id": 145,
  "from_sequence": 0,
  "steps": [...],
  "results": {
    "files_count": 3,
    "files": [
      {
        "file_name": "report.json",
        "file_key": "results/550e8400.../report.json",
        "size": 15234,
        "last_modified": "2024-12-15T10:35:00.000000+00:00",
        "download_url": "https://bucket.s3.amazonaws.com/results/550e8400.../report.json?..."
      },
      {
        "file_name": "summary.md",
        "file_key": "results/550e8400.../summary.md",
        "size": 4521,
        "last_modified": "2024-12-15T10:35:00.000000+00:00",
        "download_url": "https://bucket.s3.amazonaws.com/results/550e8400.../summary.md?..."
      },
      {
        "file_name": "screenshots.zip",
        "file_key": "results/550e8400.../screenshots.zip",
        "size": 2548000,
        "last_modified": "2024-12-15T10:35:00.000000+00:00",
        "download_url": "https://bucket.s3.amazonaws.com/results/550e8400.../screenshots.zip?..."
      }
    ],
    "s3_location": "s3://bucket-name/results/550e8400.../"
  }
}

Response Fields

Field	Type	Description
`job_id`	string	Job ID
`target_url`	string	URL being scanned
`status`	string	Current status
`effort`	string	Effort level
`created_at`	string	Job creation timestamp
`updated_at`	string	Last update timestamp
`current_step`	integer	Current agent step (1 to max_steps)
`total_sequences`	integer	Total content blocks generated
`last_sequence_id`	integer	ID of last sequence (for pagination)
`from_sequence`	integer	Requested from_step value
`steps`	array	Content blocks (text, tool_use, tool_result)
`results`	object	Result files (only when COMPLETED)
`error_message`	string	Error details (only when FAILED)

Step Types

Type	Description
`text`	Agent’s reasoning/analysis text
`tool_use`	Tool invocation (screenshot, click, type, etc.)
`tool_result`	Result of tool execution
`thinking`	Internal reasoning (when available)

Job Statuses

Status	Description
`PENDING`	Job submitted, waiting in queue
`PROCESSING`	Agent is actively scanning the website
`COMPLETED`	Scan finished successfully, results available
`FAILED`	Scan failed (see `error_message` for details)

Error Responses

400 Bad Request - Missing URL

{
  "error": "Missing required field: url"
}

400 Bad Request - Invalid URL

{
  "error": "URL must start with http:// or https://"
}

400 Bad Request - Invalid Effort

{
  "error": "Invalid effort level: extreme. Valid options are: low, medium, high"
}

403 Forbidden - Permission Denied

{
  "error": "Permission denied",
  "message": "Required permission: cua:create"
}

404 Not Found

{
  "error": "Job not found"
}

Usage Examples

Python - Complete Workflow

import requests
import time

BASE_URL = "https://stg.kyc.legaltalent.ai"
TOKEN = "YOUR_TOKEN"

headers = {
    "Authorization": f"Bearer {TOKEN}",
    "Content-Type": "application/json"
}

def submit_scan(url: str, effort: str = "medium") -> dict:
    """Submit a new CUA scan job."""
    response = requests.post(
        f"{BASE_URL}/kyc/cua-scan",
        headers=headers,
        json={"url": url, "effort": effort}
    )
    response.raise_for_status()
    return response.json()

def get_status(job_id: str, from_step: int = 0) -> dict:
    """Get job status with optional pagination."""
    params = {"from_step": from_step} if from_step > 0 else {}
    response = requests.get(
        f"{BASE_URL}/kyc/cua-scan/{job_id}",
        headers=headers,
        params=params
    )
    response.raise_for_status()
    return response.json()

def poll_until_complete(job_id: str, interval: int = 5) -> dict:
    """Poll job status until completed or failed."""
    last_sequence = 0
    
    while True:
        status = get_status(job_id, from_step=last_sequence)
        
        print(f"Status: {status['status']}, Step: {status['current_step']}")
        
        # Print new steps
        for step in status.get('steps', []):
            if step['type'] == 'text':
                print(f"  [{step['agent_step']}] {step['content'][:100]}...")
        
        if status['status'] == 'COMPLETED':
            print(f"\n✅ Scan completed! {status['results']['files_count']} result files available.")
            return status
        elif status['status'] == 'FAILED':
            print(f"\n❌ Scan failed: {status.get('error_message', 'Unknown error')}")
            return status
        
        # Update pagination cursor
        last_sequence = status.get('last_sequence_id', 0) + 1
        
        time.sleep(interval)

# Example usage
if __name__ == "__main__":
    # Submit scan
    job = submit_scan("https://suspicious-shop.com", effort="medium")
    print(f"Job submitted: {job['job_id']}")
    
    # Poll until complete
    result = poll_until_complete(job['job_id'])
    
    # Download results
    if result['status'] == 'COMPLETED':
        for file in result['results']['files']:
            print(f"Download: {file['file_name']} -> {file['download_url']}")

JavaScript - Real-time Updates

const BASE_URL = 'https://stg.kyc.legaltalent.ai';
const TOKEN = 'YOUR_TOKEN';

const headers = {
  'Authorization': `Bearer ${TOKEN}`,
  'Content-Type': 'application/json'
};

async function submitScan(url, effort = 'medium') {
  const response = await fetch(`${BASE_URL}/kyc/cua-scan`, {
    method: 'POST',
    headers,
    body: JSON.stringify({ url, effort })
  });
  return response.json();
}

async function getStatus(jobId, fromStep = 0) {
  const url = fromStep > 0 
    ? `${BASE_URL}/kyc/cua-scan/${jobId}?from_step=${fromStep}`
    : `${BASE_URL}/kyc/cua-scan/${jobId}`;
  
  const response = await fetch(url, { headers });
  return response.json();
}

async function pollWithCallback(jobId, onUpdate, interval = 3000) {
  let lastSequence = 0;
  
  const poll = async () => {
    const status = await getStatus(jobId, lastSequence);
    
    // Notify callback with new data
    onUpdate(status);
    
    if (status.status === 'COMPLETED' || status.status === 'FAILED') {
      return status;
    }
    
    // Update cursor for next poll
    lastSequence = (status.last_sequence_id || 0) + 1;
    
    // Schedule next poll
    await new Promise(resolve => setTimeout(resolve, interval));
    return poll();
  };
  
  return poll();
}

// Example: Real-time UI updates
async function scanWithLiveUpdates(url) {
  const job = await submitScan(url, 'medium');
  console.log(`Job started: ${job.job_id}`);
  
  const result = await pollWithCallback(job.job_id, (status) => {
    console.log(`[${status.status}] Step ${status.current_step}`);
    
    // Update UI with new steps
    status.steps?.forEach(step => {
      if (step.type === 'text') {
        console.log(`  Agent: ${step.content.slice(0, 80)}...`);
      } else if (step.type === 'tool_result' && step.screenshot_url) {
        console.log(`  📸 Screenshot: ${step.screenshot_url}`);
      }
    });
  });
  
  if (result.status === 'COMPLETED') {
    console.log('\n✅ Scan complete!');
    console.log('Results:', result.results.files.map(f => f.file_name));
  }
  
  return result;
}

List Jobs with Filtering

def list_jobs(status=None, limit=20, sort='desc'):
    """List scan jobs with optional filtering."""
    params = {'limit': limit, 'sort': sort}
    if status:
        params['status'] = status
    
    response = requests.get(
        f"{BASE_URL}/kyc/cua-scan",
        headers=headers,
        params=params
    )
    return response.json()

def list_all_jobs(status=None):
    """Iterate through all jobs using pagination."""
    cursor = None
    all_jobs = []
    
    while True:
        params = {'limit': 100}
        if status:
            params['status'] = status
        if cursor:
            params['cursor'] = cursor
        
        response = requests.get(
            f"{BASE_URL}/kyc/cua-scan",
            headers=headers,
            params=params
        )
        data = response.json()
        
        all_jobs.extend(data['items'])
        
        if not data['has_more']:
            break
        
        cursor = data['cursor']
    
    return all_jobs

# Get all completed jobs
completed_jobs = list_all_jobs(status='COMPLETED')
print(f"Total completed scans: {len(completed_jobs)}")

Best Practices

Choosing Effort Level

Low (20 steps): Quick checks, simple landing pages
Medium (50 steps): Standard analysis, e-commerce sites
High (80 steps): Deep investigation, complex sites with multiple pages

Efficient Polling

Use from_step parameter to only fetch new content
Poll every 3-5 seconds during active processing
Stop polling when status is COMPLETED or FAILED

Rate Limiting

Maximum 2 concurrent jobs per tenant
Maximum 10 concurrent jobs globally
Jobs exceeding limits are queued automatically

Result Files

report.json: Structured analysis data
summary.md: Human-readable summary
screenshots.zip: All captured screenshots
Download URLs expire after 1 hour

Error Handling

Implement retry logic for transient failures
Check error_message field when status is FAILED
Store job_id for later retrieval of results

Overview

Compliance Screening

AI Analysis

Batch Processing

KYC Workflows & Sessions

Tenant Management

Watchlists

Reporting & Analytics

Monitoring

​Overview

​Endpoints

​Authentication

​Submit Scan Job

​Request Parameters

​Effort Levels

​Request Example

​Response

​Response Fields

​List Jobs

​Query Parameters

​Request Example

​Response

​Response Fields

​Get Job Status

​Path Parameters

​Query Parameters

​Request Example

​Response - Processing

​Response - Completed

​Response Fields

​Step Types

​Job Statuses

​Error Responses

​400 Bad Request - Missing URL

​400 Bad Request - Invalid URL

​400 Bad Request - Invalid Effort

​403 Forbidden - Permission Denied

​404 Not Found

​Usage Examples

​Python - Complete Workflow

​JavaScript - Real-time Updates

​List Jobs with Filtering

​Best Practices

​Choosing Effort Level

​Efficient Polling

​Rate Limiting

​Result Files

​Error Handling

Overview

Endpoints

Authentication

Submit Scan Job

Request Parameters

Effort Levels

Request Example

Response

Response Fields

List Jobs

Query Parameters

Request Example

Response

Response Fields

Get Job Status

Path Parameters

Query Parameters

Request Example

Response - Processing

Response - Completed

Response Fields

Step Types

Job Statuses

Error Responses

400 Bad Request - Missing URL

400 Bad Request - Invalid URL

400 Bad Request - Invalid Effort

403 Forbidden - Permission Denied

404 Not Found

Usage Examples

Python - Complete Workflow

JavaScript - Real-time Updates

List Jobs with Filtering

Best Practices

Choosing Effort Level

Efficient Polling

Rate Limiting

Result Files

Error Handling