Trove

{
  "service_type": "platform",
  "base_url": "https://trove.nla.gov.au",
  "auth_method": "none",
  "auth_config": {},
  "endpoints": [],
  "pricing_model": {
    "type": "free",
    "details": {
      "access": "free to access and use for personal, research, and educational purposes. The API is also free but may require a key for higher usage.",
      "restrictions": "Bulk scraping or automated access may be limited by Anubis anti-bot protection."
    }
  },
  "rate_limits": {
    "api": "Unclear, but likely similar to the old Trove API limits (e.g., 10,000 requests per day).",
    "web": "Subject to Anubis proof-of-work challenge during bot detection."
  },
  "capabilities": [
    "Search across digitised newspapers from Australia",
    "Full text search of historical newspapers",
    "Access to OCR text and article metadata",
    "Metadata including title, date, and page",
    "Access to PDF or image downloads of newspaper pages",
    "Zoning and article identification",
    "Search filters by date, state, title, and category",
    "API endpoints for search and retrieval (legacy v2 API documented on the site)",
    "Integration with Trove's catalogue and other NLA collections"
  ],
  "raw_analysis": "Trove is a platform by the National Library of Australia (NLA) that provides access to millions of digitised newspapers, books, images, and other cultural heritage materials. It is primarily used by researchers, historians, genealogists, and the general public to search historical newspapers. The platform has a mature web interface and a well-documented API (v2) that allows programmatic access to newspaper articles, including OCR text and metadata. However, the current API documentation page is partially blocked by Anubis, a JavaScript anti-bot proof-of-work challenge, which suggests that the site is taking measures to prevent automated scraping. The API itself is free and offers rich data for research, but the anti-bot challenge could hinder extensive automated access. Key capabilities include full-text search, filtering, and direct access to article-level data. There are integrations with tools like Python libraries (e.g., `troveharvester` can be used by researchers) and it is widely used in digital humanities. The platform is mature, having been operational for over a decade, and its API is stable, though the future direction may see changes. No explicit pricing model is mentioned for the API; it appears to be free, but rate limits are not formally stated. Overall, Trove is a valuable resource for accessing historical newspaper content, but automated use may require navigating the Anubis protection."
}

3/3 tests passed

{
  "overall": 44,
  "dimensions": {
    "token_efficiency": 8.0,
    "first_try_success": 6.0,
    "response_parseability": 6.0,
    "error_clarity": 4.0,
    "doc_quality": 7.0,
    "auth_simplicity": 5.0,
    "latency": 6.0,
    "consistency": 5.0
  },
  "pricing_normalized": {
    "type": "free",
    "usage_limits": "Free for personal, research, educational; API free but may need key for higher usage; bulk scraping limited by Anubis anti-bot"
  },
  "issues": [
    "Anti-bot protection (Anubis) serves a challenge page, blocking automated access and potentially confusing agents. The website response returned a 'Making sure you're not a bot!' page, which severely hampers direct scraping or automated interactions.",
    "The platform lacks an llms.txt file (the endpoint returned the same anti-bot page instead of structured AI-friendly documentation), making it harder for agents to discover and use the platform.",
    "The legacy v2 API is documented but may require a key and additional steps for higher usage, adding friction for quick onboarding."
  ],
  "recommendations": [
    "Provide an llms.txt or structured API documentation that is accessible without passing the anti-bot challenge, to facilitate agent integration.",
    "Offer a clear self-service API key signup process with instant approval for low-volume usage, reducing onboarding friction.",
    "Explicitly document rate limits and acceptable use policies for automated access, so agents can operate within guidelines without hitting blocks."
  ]
}