Wellness Project

Wellness Project puts your training, nutrition, sleep, recovery, labs, and body composition in one place, then gives you a bench of eight specialist AI advisors that read your real logged data and tell you what to do next. Most health apps are silos. Your ring knows your sleep, your gym app knows your lifts, your food tracker knows your macros, and nothing knows all three. The advice stays generic because each app sees a tenth of the picture. This one connects them. WHAT IT TRACKS - Workouts, sets, reps, and exercise progression over time - Nutrition, calories, and macros against your targets - Sleep stages, HRV, resting heart rate, and recovery - Body weight, composition, and trend lines - Blood work and lab markers - Supplements, injuries, and daily wellbeing check-ins - Runs, steps, and cardio WEARABLES AND DEVICES IT SYNCS Fitbit and Oura Ring connect directly, no phone required. Everything else arrives through Apple Health on iOS or Google Health Connect on Android, which means Apple Watch, Garmin, Whoop, Samsung Galaxy Watch, Pixel Watch, Polar, Coros, Amazfit, Zepp, Peloton, Strava, and Wahoo all land in the same dashboard as long as their app writes to the platform health store. Smart scales too: Withings, Renpho, Eufy, and Wyze push weight and body fat the same way. Pick any hardware. The app reads it either way. EIGHT SPECIALIST ADVISORS, NOT ONE CHATBOT Instead of a single general assistant, you get a panel: strength, nutrition, longevity, biohacking, hypertrophy, running, protocol design, and injury. Each one reads your actual logged and synced data before answering. Ask why your bench stalled and you get an answer built on your last twelve weeks of sets, your protein intake, and your sleep debt. Not a listicle. They also disagree. The longevity advisor and the hypertrophy advisor want different things from your week, and seeing that tension against your own numbers beats one confident voice pretending the tradeoff isn't there. IT WORKS INSIDE YOUR AI ASSISTANT Wellness Project runs an OAuth-protected MCP server: https://wellnessproject.ai/api/mcp Add it as a custom connector in Claude (claude.ai and Claude Desktop, with native Claude widgets), ChatGPT, Grok, or Mistral Le Chat. Any MCP-capable client works, those four are the ones verified end to end. Once connected you can log a workout, ask a coach a question, or pull up your sleep trend from inside the assistant you already have open. Your data reaches the model only when it asks, in that conversation, never as a batch export. The whole product is built as structured, machine-readable data first. The app is one client. Your assistant is another. LOGGING THAT DOESN'T FEEL LIKE DATA ENTRY - Chat logging: describe a meal or a workout in plain language and it lands in your log, parsed - Photo meal scan: pick photos from your camera roll and the app finds the meals to log in one tap - Twelve languages THE FIT SCORE One number, updated daily, combining training load, recovery, sleep, and nutrition. It is a signal, not a scoreboard. Tap it and you see exactly which component moved and why, so you can act on it instead of guessing. It is deliberately hard to game. HOW IT COMPARES Against MyFitnessPal, Cronometer, MacroFactor, Lose It, and Noom: those track food well and stop there. None of them see your training load, your HRV, or your bloodwork, so none can tell you your cut is failing because you are underslept. Against Whoop and Oura: excellent sensors, closed loops. They score your recovery but never learn what you ate or lifted. Wellness Project reads their output and adds the other half. Against Strava and Fitbit Premium: strong at one activity, silent on the rest. Against asking ChatGPT directly: the model has no memory of your last twelve weeks. The MCP connector gives it that memory. A NORMAL DAY WITH IT Morning: last night's sleep and HRV synced from your ring before you opened anything. Lunch: photograph the plate, tap once, macros logged. Gym: you