Materials Platform for Data Science

{
  "service_type": "platform",
  "base_url": "https://mpds.io",
  "auth_method": "api_key",
  "auth_config": { "api_key_header": "Authorization", "api_key_param": "apikey" },
  "endpoints": [
    { "path": "/api/v1/structures", "method": "POST", "description": "Query for crystal structure data" },
    { "path": "/api/v1/phases", "method": "POST", "description": "Query for phase diagram data" },
    { "path": "/api/v1/properties", "method": "POST", "description": "Query for physical/chemical properties" },
    { "path": "/api/v1/systems", "method": "POST", "description": "Query for chemical systems" },
    { "path": "/api/v1/query", "method": "POST", "description": "General-purpose query endpoint" }
  ],
  "pricing_model": { "type": "freemium", "details": { "free_tier": "Limited free access with API key", "subscription_tier": "Full access for academic/research institutions" } },
  "rate_limits": {},
  "capabilities": [
    "Serve curated experimental data for materials science (crystallography, thermodynamics, properties)",
    "Provide REST-like API endpoints for programmatic access",
    "Support queries by chemical system, phase, structure, or property",
    "Enable integration with data science pipelines in materials informatics",
    "Offer dataset download and export for offline analysis"
  ],
  "raw_analysis": "MPDS (Materials Platform for Data Science) is a specialized online database and API for materials science, focusing on curated experimental data extracted from primary literature. The service is designed for researchers, data scientists, and engineers in materials discovery and design. It provides access to crystallographic structures (including symmetry, lattice parameters, atomic positions), phase diagrams (including stability regions and reactions), and physicochemical properties (e.g., thermal, electrical, magnetic). The platform is mature, having been developed over several years (since about 2019) with backing from major research institutions. The API is RESTful, requiring an API key (likely obtained via registration or academic subscription), and supports JSON responses. It is designed as a programmatic data source for high-throughput materials informatics, offering endpoints for structures, phases, properties, and chemical systems. Integration with popular data science tools (Python, pandas) is common, though no official SDK is advertised. The platform is not open-source but provides freemium access; free limited queries are possible, while extended access requires a paid institutional subscription. There is no public documentation on rate limits, but typical academic usage is moderate. The service is relevant to materials science communities (e.g., for machine learning, thermochemistry, and crystal structure prediction)."
}

2/3 tests passed

{
  "overall": 68,
  "dimensions": {
    "token_efficiency": 8.0,
    "first_try_success": 6.0,
    "response_parseability": 8.0,
    "error_clarity": 6.0,
    "doc_quality": 6.0,
    "auth_simplicity": 7.0,
    "latency": 8.0,
    "consistency": 7.0
  },
  "pricing_normalized": {
    "free_tier": "Limited free access with API key",
    "subscription_tier": "Full access for academic/research institutions"
  },
  "issues": [
    "No llms.txt file (404) — agent discoverability compromised",
    "Pricing details unclear; requires manual investigation for exact limits",
    "Onboarding may involve API key registration which can be a bit involved"
  ],
  "recommendations": [
    "Add an llms.txt file to clearly describe API capabilities and usage",
    "Publish clear documentation on free tier limits and subscription details",
    "Consider simplifying API key acquisition with SSO or magic link"
  ]
}