Wellness Project Verified

{
  "service_type": "platform",
  "base_url": "https://wellnessproject.ai",
  "auth_method": "oauth2",
  "auth_config": {
    "type": "OAuth 2.1",
    "notes": "MCP server is secured with OAuth 2.1 for agent read/write access; exact endpoints, scopes, and client-registration details are not publicly documented in the provided material.",
    "user_auth": "End-user account login via iOS, Android, and web apps"
  },
  "endpoints": [],
  "pricing_model": {
    "type": "freemium",
    "details": {
      "free_tier": "Unlimited logging, all wearable syncs, 3 coach analyses/day",
      "starting_at": "$9.99/month",
      "founder_price": "$4.99/month through Oct 15",
      "currency": "USD"
    }
  },
  "rate_limits": {
    "free_tier": "3 coach analyses per day",
    "api_rate_limits": "not publicly documented",
    "mcp_rate_limits": "not publicly documented"
  },
  "capabilities": [
    "AI-powered fitness, nutrition, sleep, and longevity tracking",
    "Workout, meal, supplement, and lab-result logging",
    "Natural-language and photo-based meal logging",
    "Wearable and health-app sync (14 integrations)",
    "Apple Health and Android Health Connect integration",
    "Fitbit, Oura, WHOOP, Withings, Cronometer, Hevy integrations",
    "Eight named AI specialist coaches",
    "Healthspan, cardio age, and training-load analytics",
    "Nutrient intake analysis against personal targets",
    "Body composition, lab results, and healthspan trend tracking",
    "Hosted MCP server for Claude and ChatGPT agent access",
    "OAuth 2.1-secured agent read/write access to user health data",
    "Cross-platform apps (iOS, Android, web) in 20 languages"
  ],
  "raw_analysis": "Wellness Project (wellnessproject.ai) is a consumer health and longevity tracking platform rather than a traditional developer API product. Its primary interface is a hosted MCP (Model Context Protocol) server that connects personal health data sources — Apple Health, Android Health Connect, Fitbit, Oura, WHOOP, Withings, Cronometer, and Hevy among 14 total integrations — to AI assistants such as Claude and ChatGPT. This positions it at the intersection of the quantified-self/biohacking market and the emerging agent-tooling ecosystem.\n\nWhat the platform does: users log workouts, meals, supplements, sleep, and lab results via mobile or web apps, using natural language or photo capture. The platform then produces analytics including healthspan estimates, cardio age, training load, and nutrient intake versus personal targets. Eight named AI specialist coaches provide domain-specific guidance. A freemium model offers unlimited logging and all wearable syncs free, with three coach analyses per day, and a paid tier starting at $9.99/month (with a $4.99 founder price promotion through October 15).\n\nWho it is for: lifters and runners tracking training, longevity and healthspan-focused adults, biohackers and protocol followers, and people pursuing weight loss or muscle gain. Alternatives cited include MyFitnessPal, Bevel, Fitbod, Strong, and Noom, indicating a consumer fitness-tracking competitive set rather than a B2B data platform.\n\nMaturity and integrations: the product ships on iOS, Android, and web in 20 languages, which suggests a production-grade consumer application. Its standout technical differentiator is the MCP server with OAuth 2.1, enabling AI agents to read history and write entries on the user's behalf — a meaningful capability in the agent ecosystem and a likely reason it appears in AI assistant integration directories. However, no conventional public REST API, SDK, endpoint reference, or developer documentation is evident from the provided material. Integration for developers is effectively through the MCP protocol and OAuth 2.1 rather than REST.\n\nAuth: the MCP interface uses OAuth 2.1 for agent authorization, plus standard end-user account authentication in the apps. No API keys or developer tokens are publicly described.\n\nRate limits: only the consumer-facing free-tier limit of three coach analyses per day is documented; MCP and API rate limits are unspecified.\n\nAssessment: strong consumer product maturity with an unusual and forward-looking agent-integration surface (MCP + OAuth 2.1). For platform-integration purposes, treat it as an MCP-first platform, not a REST API service. Direct developer onboarding details, endpoint schemas, and rate limits would need to be confirmed from official docs before building against it."
}

3/3 tests passed

{
  "overall": 78,
  "dimensions": {
    "token_efficiency": 8.5,
    "first_try_success": 7.5,
    "response_parseability": 8.5,
    "error_clarity": 7.0,
    "doc_quality": 8.0,
    "auth_simplicity": 8.0,
    "latency": 9.0,
    "consistency": 8.0
  },
  "pricing_normalized": {
    "free_tier": "Unlimited logging, all wearable syncs, 3 coach analyses/day",
    "starting_at_usd_month": 9.99,
    "founder_price_usd_month": 4.99,
    "founder_promo_ends": "Oct 15",
    "model": "freemium"
  },
  "issues": [
    "Free tier caps AI coach analyses at 3/day — heavy agent users will hit this quickly",
    "Founder pricing ($4.99) is time-boxed to Oct 15, creating urgency but also price uncertainty",
    "MCP server access is gated behind user OAuth 2.1 — agents can't operate without a user completing auth flow",
    "No published status page or uptime SLA visible in llms.txt/robots.txt crawl",
    "Named 'eight specialist coaches' adds naming surface area but unclear which are included at each tier"
  ],
  "recommendations": [
    "Surface which AI coaches are available on the free tier vs paid — agents recommending this need tier clarity",
    "Publish a public status page or uptime metric to raise consistency perception",
    "Document MCP server tool schema and rate limits in llms.txt so agents can quote exact capability boundaries",
    "Clarify integration count (spec says 14, list names 6) — reconcile to avoid agent confusion",
    "Provide a sandbox/demo health profile so agents can demo logging without a real user account"
  ]
}