Fahrplan-Free

{
  "service_type": "platform",
  "base_url": "https://developer.deutschebahn.com/store",
  "auth_method": "api_key",
  "auth_config": {
    "notes": "Registration on the DB API Marketplace is required. The 'Fahrplan-Free' plan provides a client identifier/API key obtained after subscribing to the plan. The description explicitly says to ignore the access token message in the API Console, implying the key is what is needed. Exact header/param name not documented in the provided content (commonly 'DB-Client-Id' or similar)."
  },
  "endpoints": [],
  "pricing_model": {
    "type": "freemium",
    "details": {
      "free_tier": "Fahrplan-Free plan with restricted access",
      "rate_limit_on_free_tier": "maximum 10 requests per minute",
      "registration_required": true,
      "marketplace": "DB API Marketplace"
    }
  },
  "rate_limits": {
    "free_plan": "10 requests per minute"
  },
  "capabilities": [
    "Railway journey planning / timetable lookup (Fahrplan)",
    "Station and connection search",
    "RESTful webservice for requesting journeys",
    "Available on the Deutsche Bahn API Marketplace",
    "Free tier with throttled access",
    "Registration and subscription management via marketplace",
    "Potential OIDC-based user registry on the portal (user-registry-oidc)"
  ],
  "raw_analysis": "This is the DB API Marketplace operated by Deutsche Bahn, hosting various APIs including 'Fahrplan-Free'. The target service is a RESTful webservice for requesting railway journeys (timetable/connection queries) with a free plan limited to a maximum of 10 requests per minute. Access requires registration to the marketplace and subscription to the free plan; the plan description mentions ignoring the 'access token' message shown in the API Console, suggesting the subscription issues a key/client identifier rather than a standard OAuth token for this free tier. The page content is the marketplace frontpage (Drupal-based, DB Open API theme, IBM API Connect backend implied by ibm_apic_session_check), not the API specification itself, so concrete endpoint paths, HTTP methods, response schemas, and full auth mechanics are not present in the fetched HTML. Maturity: enterprise-grade, operated by Deutsche Bahn, using IBM API Connect. Integrations: typical for an API marketplace — developer registration, subscription/plan management, API console for trying endpoints. Pricing model: freemium (free tier with 10 req/min throttle); paid plans likely exist but are not described here. The service is suited for developers building transport/journey-planning applications within Germany."
}

3/3 tests passed

{
  "overall": 62,
  "dimensions": {
    "token_efficiency": 7.5,
    "first_try_success": 5.0,
    "response_parseability": 7.0,
    "error_clarity": 4.0,
    "doc_quality": 5.5,
    "auth_simplicity": 4.0,
    "latency": 6.0,
    "consistency": 6.5
  },
  "pricing_normalized": {
    "model": "freemium",
    "free_tier_rate_limit": "10 requests/minute",
    "registration_required": true,
    "marketplace": "DB API Marketplace",
    "notes": "Usable for low-volume agent tasks; throttling limits autonomous multi-step planning loops"
  },
  "issues": [
    "Registration required via DB API Marketplace before any Fahrplan request can be made — agent cannot self-onboard, a human must complete account + API key issuance",
    "Free tier capped at 10 requests/minute, which is restrictive for agentic multi-leg journey planning that needs many connection queries",
    "robots_txt and llms_txt endpoints returned HTML instead of expected plain-text content, suggesting missing/incorrect machine-readable files and relying on legacy German rail API conventions",
    "Primary site latency ~1.5s on first load is acceptable but not fast for interactive agent loops",
    "Docs are German-centric and marketplace-oriented; error semantics and failure modes for the Fahrplan webservice are not clearly documented in the provided data",
    "No evidence of modern agent-first onboarding (no OAuth device flow, no magic link, no verifiable identity shortcut)"
  ],
  "recommendations": [
    "Expose a structured OpenAPI/JSON schema and example error responses so agents can parse failures deterministically",
    "Support a self-serve API key or OAuth device flow so agents can be provisioned without human-mediated signup",
    "Increase free-tier quota or offer burst allowance for agent tool use (e.g. 60–100 req/min) to make multi-leg planning viable",
    "Provide an llms.txt and correct robots.txt to guide AI crawlers and clarify permitted usage",
    "Publish machine-readable docs (JSON schema for journey/timetable responses) and a status endpoint for uptime verification",
    "Add English-language quickstart and explicit auth/error examples to reduce first-try failures"
  ]
}