thirds.ai Verified

{
  "service_type": "platform",
  "base_url": "https://thirds.ai",
  "auth_method": "unknown",
  "auth_config": {},
  "endpoints": [],
  "pricing_model": {
    "type": "unknown",
    "details": {}
  },
  "rate_limits": {},
  "capabilities": [
    "AI-powered platform (unconfirmed specifics)"
  ],
  "raw_analysis": "thirds.ai appears to be a service registered under the 'ai' category via what looks like an automated benchmark URL submission process. The only available information is the domain name itself and the categorization as an AI-related service. No public documentation, API references, pricing pages, or feature descriptions were provided or discoverable from the supplied content.\n\nKey observations:\n- Domain name 'thirds.ai' — the '.ai' TLD strongly suggests an artificial intelligence product, but the 'thirds' label gives no clear indication of function (could relate to thirds, fractions, partitioning, a third-party tool, or a personal/brand name).\n- The registration note ('Auto-registered via benchmark URL submission') indicates this entry was created programmatically, not through direct product research, so it carries low confidence about the platform's actual capabilities.\n- Nothing is known about maturity (could be pre-launch, stealth, or established), target audience, integrations, SDKs, or whether any public REST API exists.\n\nWhat cannot be determined from the given input:\n- Whether thirds.ai exposes any API at all (REST, GraphQL, webhooks, etc.).\n- Authentication methods (API keys, OAuth, none).\n- Pricing model (free, freemium, subscription, enterprise-only).\n- Rate limits or usage quotas.\n- Concrete platform capabilities beyond a generic 'AI' tag.\n\nRecommendation: To characterize this platform accurately, a follow-up reconnaissance pass is required — fetch the live site, inspect /docs, /api, /pricing, /developers, check for an OpenAPI/Swagger spec, review robots.txt and sitemap.xml, and look for the company behind the domain (WHOIS, LinkedIn, product launch posts). Until then, all fields beyond base_url and the coarse 'ai' category should be treated as unknown and not relied upon."
}

3/3 tests passed

{
  "overall": 63,
  "dimensions": {
    "token_efficiency": 7.0,
    "first_try_success": 6.5,
    "response_parseability": 7.5,
    "error_clarity": 5.0,
    "doc_quality": 6.0,
    "auth_simplicity": 5.0,
    "latency": 7.5,
    "consistency": 7.0
  },
  "pricing_normalized": {
    "model": "unknown",
    "notes": "No pricing data returned by checks; llms_txt mentions API and AI tools but pricing tier not surfaced."
  },
  "issues": [
    "Auth endpoints (/signin, /oauth, /auth, /session) disallowed in robots.txt, suggesting non-trivial or gated auth flows for agents.",
    "No llms.txt pricing or rate-limit info — agents can't predict cost or throttling before integrating.",
    "API/MCP endpoints (/v1, /mcp) are robots-disallowed, weakening crawlability of integration docs for agents.",
    "Pricing model unknown — reduces confidence for recommendation without manual verification.",
    "Only 3 checks run; no error-shape, schema, or status-page evidence to judge error clarity and stability."
  ],
  "recommendations": [
    "Publish pricing and rate limits in llms.txt or a machine-readable endpoint so agents can self-assess fit.",
    "Add a /llms-full.txt or OpenAPI/JSON schema reference that is crawl-allowed to improve parseability and onboarding.",
    "Provide a status page and documented error codes/envelopes to raise error_clarity and consistency scores.",
    "Offer an agent-friendly auth path (API key via magic link / Notlogin) to simplify onboarding beyond browser signin.",
    "Expose a lightweight public health/version endpoint outside the disallow list for uptime verification."
  ]
}