An API to extract structured data from any

{
  "service_type": "platform",
  "base_url": "https://ninjadoc.ai",
  "auth_method": "api_key",
  "auth_config": {
    "header": "Authorization",
    "scheme": "Bearer"
  },
  "endpoints": [
    {
      "path": "/api/extract",
      "method": "POST",
      "description": "Extract structured data from a document URL or file upload",
      "auth_required": true
    },
    {
      "path": "/api/documents",
      "method": "GET",
      "description": "List processed documents",
      "auth_required": true
    },
    {
      "path": "/api/processors",
      "method": "GET",
      "description": "List query collections/processors",
      "auth_required": true
    }
  ],
  "pricing_model": {
    "type": "subscription",
    "details": {
      "free_tier": true,
      "paid_tiers": true,
      "credit_based": true,
      "price_range": "unknown"
    }
  },
  "rate_limits": {
    "default": "not documented"
  },
  "capabilities": [
    "extract structured data from any document without training",
    "support for multiple document formats (PDF, images, etc.)",
    "no-code configuration for data extraction",
    "query collections for organizing processors",
    "web dashboard for managing documents and processors",
    "API-first design for integration",
    "authentication via Clerk",
    "server-side rendering with React Router"
  ],
  "raw_analysis": "Ninjadoc AI is a service that provides an API to extract structured data from any document without requiring training. The platform is designed for developers and businesses that need to automate data extraction from PDFs, images, and other document types. It offers a no-code approach, allowing users to define extraction schemas without machine learning expertise. The service is relatively new (recently posted on Hacker News as 'Show HN'), indicating it is in early stages but actively developed. The tech stack uses React Router with server-side rendering, Clerk for authentication, and Google Analytics/Twitter tracking for marketing. The landing page is at / and the dashboard at /app/dashboard. The API is likely RESTful, though endpoints are not publicly documented (the provided URL returned a 404). The platform appears to offer a freemium model with subscription tiers, likely based on usage or credits. Existing integrations include Clerk for user management and Google Analytics for tracking. Overall, it's a promising tool for document data extraction, with a modern web app and a clear focus on API usability."
}

3/3 tests passed

```json
{
  "overall": 55,
  "dimensions": {
    "token_efficiency": 7.0,
    "first_try_success": 6.0,
    "response_parseability": 8.0,
    "error_clarity": 5.0,
    "doc_quality": 5.0,
    "auth_simplicity": 6.0,
    "latency": 5.5,
    "consistency": 5.0
  },
  "pricing_normalized": {
    "free_tier": true,
    "credit_based": true,
    "subscription": true,
    "price_range": "unknown"
  },
  "issues": [
    "Pricing details are unknown — agents cannot estimate costs for users",
    "llms.txt endpoint returns HTML instead of plain text (failed machine-readable format)",
    "No security headers detected on the website",
    "Search engine crawls may be slowed by explicit crawl delay in robots.txt",
    "Website latency at 1325ms exceeds the 1-second threshold for optimal agent experience"
  ],
  "recommendations": [
    "Publish a clear llms.txt with plain-text markdown for agent consumption",
    "Add transparent pricing page to reduce uncertainty during recommendation",
    "Improve latency to under 800ms for better agent response times",
    "Add security headers to improve trust signals",
    "Provide API example responses for common extraction use cases"
  ]
}
```