Archive API

{
  "service_type": "platform",
  "base_url": "https://developer.nytimes.com",
  "auth_method": "api_key",
  "auth_config": {
    "param_name": "api-key",
    "registration_url": "https://developer.nytimes.com/get-started",
    "notes": "Requires a free NYT Developer account and an API key issued per-application. Key passed as the 'api-key' query parameter on every request."
  },
  "endpoints": [
    {
      "name": "Archive by Month",
      "method": "GET",
      "path": "/svc/archive/v1/{year}/{month}.json",
      "host": "https://api.nytimes.com",
      "description": "Returns a list of article metadata (title, abstract, url, byline, published_date, section, etc.) for all NYT articles published in the given year/month.",
      "params": {
        "year": "integer, 1851..present",
        "month": "integer, 1..12",
        "api-key": "required"
      },
      "example": "https://api.nytimes.com/svc/archive/v1/2024/1.json?api-key=YOUR_KEY"
    }
  ],
  "pricing_model": {
    "type": "freemium",
    "details": {
      "free_tier": "Available with rate limits (typically 10 requests/minute, 4,000 requests/day for the free tier)",
      "paid_tiers": "Higher volume plans available via NYT Developer portal",
      "notes": "No cost for typical metadata-cataloging use cases within rate limits"
    }
  },
  "rate_limits": {
    "free_tier": {
      "requests_per_minute": 10,
      "requests_per_day": 4000
    },
    "notes": "Rate limits vary by API key tier; see https://developer.nytimes.com/faq for current figures"
  },
  "capabilities": [
    "Monthly article metadata retrieval going back to 1851",
    "Article title, abstract, URL, byline, section, subsection, published date",
    "Bulk corpus construction for local databases / datasets",
    "Historical coverage spanning 170+ years",
    "JSON responses suitable for ETL pipelines",
    "Data journalism, NLP, and archival research support"
  ],
  "raw_analysis": "The NYT Archive API is one of several APIs offered via the New York Times Developer Network (https://developer.nytimes.com/), alongside Article Search, Top Stories, Books, Movie Reviews, and others. It is a REST/JSON HTTP API hosted at api.nytimes.com, NOT the nytimes.com domain in the provided URL — the given URL 'https://nytimes.com:archive' appears to be a malformed or shorthand reference. The actual endpoint pattern is /svc/archive/v1/{year}/{month}.json.\n\nWhat it does: Returns a paginated-style list of article metadata (not full text) for every article published by the NYT in a given month, from 1851 to the present. This makes it uniquely valuable for reconstructing a historical corpus of article metadata programmatically.\n\nWho it's for: Data journalists, researchers, NLP/ML practitioners building news-corpus datasets, digital humanities scholars, and developers building archival/database products. It is explicitly positioned for building local databases of NYT article metadata.\n\nMaturity: Very mature -- part of the long-standing NYT Developer portal, actively maintained, well-documented. This is a production-grade API backed by the NYT's internal archives.\n\nAuth: API key (query parameter 'api-key'), issued free upon registration. No OAuth, no paid-only access for basic use.\n\nLimitations: Metadata only (no article body text); one month per request means 12 requests/year for a full corpus; rate limits constrain backfilling the full 1851-present archive quickly. Terms of service restrict commercial reuse and redistribution of NYT content.\n\nIntegrations: None pre-built; consumed directly over HTTP by any client (Python requests, curl, etc.). Common integrations are custom ETL scripts and data-science notebooks. No SDK is officially published, though community Python wrappers exist.\n\nPricing: Freemium -- free with rate limits, higher-volume access available via paid agreements."
}

3/3 tests passed

{
  "overall": 71,
  "dimensions": {
    "token_efficiency": 9.0,
    "first_try_success": 6.5,
    "response_parseability": 9.0,
    "error_clarity": 7.0,
    "doc_quality": 8.0,
    "auth_simplicity": 4.5,
    "latency": 9.5,
    "consistency": 8.0
  },
  "pricing_normalized": {
    "type": "freemium",
    "free_tier": "Rate-limited (approx. 10 req/min, 4,000 req/day)",
    "paid_tiers": "Higher volume via NYT Developer portal",
    "cost_for_typical_metadata_cataloging": "Included free within rate limits"
  },
  "issues": [
    "API key registration via NYT Developer portal requires signing up through nytimes.com — agents cannot self-onboard without a human or existing NYT account.",
    "No llms.txt (endpoint returns HTML, not structured agent guidance).",
    "Free tier rate limits (10 req/min, 4,000 req/day) are restrictive for bulk corpus construction.",
    "Robots.txt only exposes a sitemap; no explicit crawling/agent usage guidance.",
    "No machine-readable API spec (OpenAPI/JSON schema) referenced in fetched endpoints."
  ],
  "recommendations": [
    "Publish an llms.txt and/or OpenAPI spec so agents can self-discover endpoints, auth, and schemas.",
    "Support simple magic-link or token-exchange onboarding so agents can provision keys without a full NYT account flow.",
    "Document rate-limit headers and quota endpoints so agents can self-throttle and avoid 429s.",
    "Provide explicit agent/crawler usage guidance in robots.txt and docs.",
    "Expose clear error schemas for common failures (invalid date range, rate limit, invalid key)."
  ]
}