Mypenny API

{
  "service_type": "platform",
  "base_url": "https://app.mypenny.ai",
  "auth_method": "unknown",
  "auth_config": {},
  "endpoints": [
    {
      "path": "/mcp",
      "method": "unknown",
      "description": "Provided as the API/entry URL; '/mcp' strongly suggests a Model Context Protocol (MCP) endpoint rather than a conventional REST API."
    }
  ],
  "pricing_model": {
    "type": "unknown",
    "details": {}
  },
  "rate_limits": {},
  "capabilities": [
    "Likely MCP (Model Context Protocol) server endpoint for AI agent/tool integration",
    "AI-related service (self-classified as 'ai')",
    "Potential natural-language or agent-driven interaction surface",
    "Possibly finance/expense-adjacent given the 'penny' naming (unverified)"
  ],
  "raw_analysis": "Mypenny (app.mypenny.ai) is an AI-category platform reachable at https://app.mypenny.ai, with the submitted API entry point being https://app.mypenny.ai/mcp. The '/mcp' path is a strong signal that this is not a traditional public REST API but rather a Model Context Protocol server — i.e., an interface designed for AI clients/agents (such as Claude Desktop, IDEs, or agent frameworks) to connect and invoke tools, rather than a documented REST surface with JSON endpoints.\n\nMaturity and public footprint: the submission appears to be auto-registered via a benchmark URL submission, with no vendor-provided documentation, OpenAPI spec, or descriptive metadata included. There is no evidence here of a documented public REST API, published auth scheme (API keys, OAuth), SDKs, or rate-limit policy. Authentication for MCP endpoints is frequently OAuth-based or key-based, but nothing in the provided content confirms this, so it is marked unknown.\n\nWhat it likely does: the name 'penny' plus the 'ai' category suggests a finance/expense/bookkeeping or personal-finance assistant, but this is inference, not confirmed by the supplied text. The only concrete technical fact is the '/mcp' endpoint, which implies agent/tool integration capabilities rather than classic developer-facing REST integration.\n\nWho it's for: plausibly end users and AI-agent builders seeking to connect an assistant to financial or productivity workflows; however, target audience, feature set, integrations, and pricing are all unverified from the given material.\n\nRecommendation: treat as a low-confidence entry. To analyze properly, retrieve any MCP server manifest or well-known metadata from https://app.mypenny.ai (e.g., /.well-known/ or MCP discovery documents), check for an OAuth authorization-server metadata document, and confirm whether a REST API exists separately from the MCP endpoint. Until then, endpoints, auth, pricing, and rate limits should be considered unknown."
}

3/3 tests passed

{
  "overall": 47,
  "dimensions": {
    "token_efficiency": 5.0,
    "first_try_success": 7.0,
    "response_parseability": 4.0,
    "error_clarity": 3.0,
    "doc_quality": 3.0,
    "auth_simplicity": 5.0,
    "latency": 9.0,
    "consistency": 6.0
  },
  "pricing_normalized": {
    "model": "unknown",
    "notes": "No pricing information available; platform appears early-stage or stealth with no publicly documented plans"
  },
  "issues": [
    "No documentation, API reference, or developer onboarding surface detected",
    "llms.txt and robots.txt return the SPA index page instead of actual files — no machine-readable discovery for agents",
    "Value prop ('secure, portable AI memory') is marketing-level; no concrete capability, format, or integration details",
    "Auth/onboarding flow cannot be assessed and likely requires a human signup — no magic link or agent-friendly auth signal",
    "No status page, changelog, or error documentation to assess reliability or failure modes",
    "Pricing and data portability/export format are undisclosed, making agent recommendations risky"
  ],
  "recommendations": [
    "Publish a real llms.txt and robots.txt (not the SPA fallback) describing capabilities, endpoints, and auth",
    "Expose a documented API or MCP server spec with request/response schemas for agent consumption",
    "Add a pricing page with a free tier and clear limits so agents can reason about cost",
    "Provide magic-link or notlogin.com-based onboarding so agents can complete setup without a human",
    "Document data model, export/portability formats, and memory retention semantics",
    "Add a status page and error code reference to improve error clarity and consistency scoring",
    "Clarify positioning: is this a memory store, an MCP service, a finance tool? Ambiguity in 'penny' naming hurts agent recommendation confidence"
  ]
}