BestYou Health Intelligence

BestYou Health Intelligence

BestYou brings real health data from your wearables into your AI agent through MCP. It pulls fitness, nutrition, sleep, and recovery signals from Apple Watch, Oura Ring, CGM sensors, and HealthKit, then coordinates them through 19 specialized on-device agents. The result is an adaptive daily action plan that responds to your actual readiness, not generic templates. Query your readiness score, generate workouts matched to your recovery state, analyze meals against your macro targets, and manage your plan through natural conversation.

Privacy note: No health data leaves your device without your explicit consent. BestYou shares processed insights only. Raw HealthKit readings stay on the iPhone. See references/security.md for details.

What Makes This Different

Most health integrations export static data or generate generic prompts. BestYou connects to a live API backed by 19 coordinated AI agents running on-device.

Setup

1. Get your API key

Open the BestYou iOS app: More → Connected Apps → OpenClaw → Generate Key

Your key looks like by_mcp_live_.... Full walkthrough: bestyou.ai/openclaw-setup

2. Set the environment variable

export BESTYOU_API_KEY="YOUR_KEY_HERE"

3. Connect via mcporter

If you don't have mcporter (OpenClaw's MCP client) installed yet, you can install it with:

npm install -g mcporter

Add BestYou as an MCP server:

mcporter config add bestyou \
  --url https://mcp.bestyou.ai/mcp \
  --header "Authorization: Bearer $BESTYOU_API_KEY"

Verify with mcporter call bestyou.get_account_link_status.

For a more detailed setup walkthrough, see references/setup.md.

Agent Behavior

Session Start

Always call get_account_link_status first. If the account is not linked or the subscription is inactive, tell the user what to fix (re-link in BestYou app, check subscription) and stop. Do not call other tools until status is confirmed.

Daily Coaching Flow

1. Call get_daily_briefing with today's date to pull the narrative, readiness score, and recommended actions. 2. Summarize in 2-3 sentences: how the user is doing, what stands out. 3. Recommend 1-2 concrete actions from the briefing.

Plan Management

Fitness

Nutrition

Progress and Trends

Style

Direct, calm, practical. You're a coach, not a cheerleader. State what matters, skip the filler.

Safety

Tool Reference

| Tool | When to use | Required parameters | |------|-------------|---------------------| | get_account_link_status | First call every session. Confirms link, subscription, scopes. | (none) | | get_daily_briefing | User asks "how am I doing today?" or starts a health conversation. | date (YYYY-MM-DD) | | get_todays_action_plan | User wants to see today's plan or scheduled blocks. | date (optional) | | get_progress_snapshot | User asks "how am I doing overall?" or wants domain scores. | date (YYYY-MM-DD) | | get_weekly_summary | End-of-week review or "how was my week?" | weekEndDate (optional) | | generate_workout | User wants a new workout built from scratch. | type, duration, equipment, experienceLevel | | analyze_meal_text | User describes food they ate for nutritional analysis. | description |

generate_workout parameters

Required: type (strength/mobility/cardio/mixed/hiit), duration (minutes), equipment (array of strings), experienceLevel (Beginner/Intermediate/Advanced)

Optional: bodyParts, targetAreas, goal, injuries, includeWarmup, warmupStyle (dynamic/activation), warmupDuration, includeCooldown, cooldownStyle (static/foam_rolling/combined), cooldownDuration

Example Conversations

Morning check-in

> User: "What's my day look like?" > > Agent: Call get_account_link_status, then get_daily_briefing with today's date. Summarize readiness, highlight top priorities, suggest 1-2 actions.

Logging a meal

> User: "Just had a chicken burrito bowl with brown rice and guac" > > Agent: Call analyze_meal_text with description="chicken burrito bowl with brown rice and guacamole". Report macros and how it fits the day's targets.

Progress check

> User: "How am I doing overall?" > > Agent: Call get_progress_snapshot with today's date. Highlight strongest and weakest domains, surface top recommendation.

Weekly review

> User: "How was my week?" > > Agent: Call get_weekly_summary. Highlight trends (improving/declining domains), top achievement, and next week's focus areas.

Quick workout

> User: "Give me a 20-minute bodyweight workout focused on strength" > > Agent: Call generate_workout with type="strength", duration=20, equipment=["bodyweight"], experienceLevel="Intermediate". Present the workout clearly with exercises, sets, reps, and rest periods.

Notes

1. Get your API key

Open the BestYou iOS app: More → Connected Apps → OpenClaw → Generate Key

Your key looks like by_mcp_live_.... Full walkthrough: bestyou.ai/openclaw-setup

2. Set the environment variable

export BESTYOU_API_KEY="YOUR_KEY_HERE"

3. Connect via mcporter

If you don't have mcporter (OpenClaw's MCP client) installed yet, you can install it with:

npm install -g mcporter

Add BestYou as an MCP server:

mcporter config add bestyou \
  --url https://mcp.bestyou.ai/mcp \
  --header "Authorization: Bearer $BESTYOU_API_KEY"

Verify with mcporter call bestyou.get_account_link_status.

For a more detailed setup walkthrough, see references/setup.md.