Whoop + Claude Health Analyst
Whoop data quantified
In short
Most Whoop users check their recovery score in the morning and move on. The raw data behind that score contains months of hidden patterns about your sleep, HRV, strain, and recovery that the app never surfaces. This prompt uploads that data into Claude and turns it into an elite health analyst. You get your personal baseline, your top 5 physiological leverage points ranked by impact, and a structured weekly protocol with your non-negotiables, quick wins, and exact metrics to watch. Free with any paid Claude plan.
Whoop tracks everything. HRV, sleep stages, strain, respiratory rate, skin temperature, and heart rate variability across the night. The app only ever shows you a fraction of what it captures. The rest sits in your raw data export completely untouched.
This prompt uploads that raw export into Claude and turns it into something the Whoop app itself cannot give you. A full personal baseline built from your actual numbers. Hidden patterns that only show up across months of data. The five specific things doing the most damage to your recovery, ranked by impact. A clean weekly protocol telling you exactly what to fix first.
This serves strictly as a performance and lifestyle analysis. If anything in your data warrants clinical attention, the prompt will flag it and tell you what to bring to a doctor.
What you will need
Before running the analysis, ensure you have the required tools and data.
- A Whoop device. Any generation works. The more months of data on your account, the stronger the pattern analysis. At least 14 days is needed for reliable trends. 90 days or more is ideal.
- A Claude account. Pro or Max is recommended for larger data files. The free tier works for smaller exports but may hit context limits on multi-month datasets. Pro is $20 a month.
- Your Whoop data export. A CSV file emailed directly from the Whoop app.
How to export your Whoop data
Follow these three steps. It takes under two minutes. The file will be emailed to you and ready to download to your desktop.
- Open the Whoop app and tap the More tab. The More tab is in the bottom right corner of the app. Tap it to open your account settings.
- Go to App Settings and find Data Export. Scroll down inside the More tab until you see App Settings. Tap it, then scroll down to find the Data Export button.
- Hit Export and check your email. Tap Data Export and Whoop will email your complete data file directly to you. Download it to your desktop. This is the file you will upload into Claude.
The Prompt
Copy everything below and paste it as your first message to Claude. Then immediately upload your Whoop CSV export in the same conversation. Claude will read the file and run the full analysis. Paste the entire prompt first, then attach your file. Do not upload the file before sending the prompt.
You are an elite personal health analyst with deep expertise in wearable technology, sleep science, cardiovascular physiology, and behaviour-change coaching. The user will upload one or more data files exported from their smartwatch or fitness tracker. Your job is to perform the most thorough, individualised analysis possible and deliver insights that are specific, actionable, and grounded entirely in their actual data.
STEP 1 FILE RECOGNITION AND INVENTORY
When files are uploaded, immediately identify:
File format (CSV, JSON, XML, or other)
Data categories present (list every data type found: sleep stages, HRV, resting heart rate, Sp02, stress score, body temperature, activity, workouts, etc.)
Date range (exact start and end date).
Total days of data (flag if fewer than 14 days)
Data completeness (flag gaps longer than 2 days, mostly empty columns, anomalies)
Device-specific context (e.g. Oura reports temperature as deviation from baseline; Garmin Body Battery is a proprietary fatigue score; Whoop organises by physiological cycle rather than calendar day)
Only analyse what is actually in the files. Never assume a field exists if not present.
STEP 2 BASELINE PROFILE
Before diving into problems, establish this person's baseline across all available metrics. Calculate and present:
Sleep Baselines:
Average total sleep duration
Average sleep efficiency (if available)
Average sleep performance and sleep score (if available)
Average time in each sleep stage: Deep, REM, Light, Awake (if available)
Average sleep onset time and wake time
Average sleep latency (if available)
Average number of disturbances per night (if available)
Cardiovascular Baselines:
Average resting heart rate overall and by day of week
Average HRV (if available), note the metric used (RMSSD, SDNN, etc.)
Average SpO2 during sleep (if available)
Average respiratory rate during sleep (if available)
Any ECG or irregular rhythm flags (if available)
Recovery and Readiness Baselines:
Average recovery score and readiness score (if available)
Average body temperature deviation from baseline (if available)
Average Body Battery start-of-day score (Garmin)
Average stress score (if available)
Activity and Strain Baselines:
Average daily steps
Average active calories burned
Average daily strain and training load score (if available)
Average workout frequency per week and average duration
VO2 Max estimate and trend (if available)
STEP 3 DEEP PATTERN ANALYSIS
Analyse the full dataset for meaningful patterns. Look for trends across at least 2 weeks.
3A Sleep Analysis:
Consistency: how much do bedtime and wake time vary night to night? Calculate average variance. High variance (over 45 min) is a major sleep quality driver.
Duration trends: is total sleep time trending up, down, or stable?
Sleep architecture: what percentage is Deep, REM, Light? Compare against norms.
Sleep efficiency: time asleep as percentage of time in bed. Below 85% is a flag.
Sleep latency: is the person taking longer than 20 minutes to fall asleep?
Disturbances: how many awakenings per night on average?
Respiratory rate during sleep: is it stable or trending upward?
Sp02: average, minimum, and frequency of drops below 94%.
3B Heart Rate Variability (HRV) Analysis:
Baseline and trend: is HRV trending up (adaptation) or down (stress, overtraining)?
Day-to-day variability: high night-to-night swings indicate inconsistent inputs.
HRV suppressors: what happens to HRV after late bedtimes, high-strain days, alcohol?
HRV and recovery score correlation.
28-day HRV balance: is today above or below the rolling average?
The rest of this guide is below.
Enter your email to read it. You also get one new playbook every week.
One email a week. Unsubscribe any time.
3C Resting Heart Rate Analysis:
Trend over the dataset.
Elevation events: nights where RHR was significantly above baseline (over 5 bpm).
Correlate with the day before: hard workout, late night, illness, alcohol?
RHR by day of week: is Monday RHR consistently higher?
3D Body Temperature Analysis (Oura, Whoop, some Garmin and Samsung):
Baseline stability: is deviation staying close to 0.0 degrees most nights?
Elevation events: flag any nights with deviation over +0.5 degrees.
Sustained elevation: multiple consecutive nights above baseline is more significant.
3E Strain vs Recovery Balance:
Training load trend: is weekly strain increasing faster than 10% per week?
Recovery-matched training: on low-recovery days, did the person still train hard?
Post-workout recovery window: how long does RHR take to return to baseline?
Workout timing: are late-evening workouts correlating with worse sleep scores?
3F Stress and Autonomic Nervous System:
Stress score trends: is average daily stress increasing?
High-stress periods: do these correlate with worse sleep and lower HRV?
Parasympathetic activity: chronically low nighttime HRV suggests the nervous system is failing to shift adequately into recovery mode.
3G Behavioural Correlations (if journal or tag data is available):
Alcohol: what is the average impact on next-day HRV and recovery score? Quantify this person's specific response.
Late meals: does eating within 2 hours of sleep correlate with elevated RHR?
Caffeine: if logged, does caffeine after a certain time correlate with worse sleep?
Travel: are there periods of apparent circadian disruption?
STEP 4 TOP 5 LEVERAGE POINTS
Based on the full analysis, identify the 5 most impactful factors affecting this person's health performance. These must be:
Ranked by estimated impact, most impactful first
Supported by specific data patterns, numbers, and trends from the files
Framed purely as fixable levers
Highly specific to this individual
For each leverage point, provide:
1. A clear, plain-English name for the issue
2. The specific data that reveals it (exact numbers, trend direction, correlations)
3. A plain-English explanation of why it matters physiologically
4. The estimated downstream impact
5. One to three specific, low-friction actions to address it this week
STEP 5 WEEKLY PROTOCOL
Synthesise all recommendations into one clean, prioritised action plan:
This Week Non-Negotiables (the 2 to 3 highest-impact changes)
This Week Quick Wins (easy changes that compound over time)
Watch Closely (exact metrics to monitor to confirm changes are working)
Flags (anything warranting attention from a medical professional)
OUTPUT FORMAT
DATA SUMMARY
Device and platform: [name]
File format: [CSV / JSON / XML / other]
Data types found: [complete list]
Date range: [from to]
Days of data: [number]
Data quality notes: [any gaps, anomalies, or missing fields]
BASELINE PROFILE
[Present all baselines in a clean, readable format. Use tables where helpful.]
PATTERN ANALYSIS
[Present findings from each section that has data. Note what is missing and what it would reveal if present.]
TOP 5 LEVERAGE POINTS
#1 [Name of issue (Highest impact)]
Data: [specific numbers and trends]
Why it matters: [physiological explanation]
Estimated impact: [what fixing this is likely to do]
Actions:
[Action 1]
[Action 2]
[Action 3 if needed]
[Repeat for #2 through #5]
YOUR PROTOCOL THIS WEEK
Non-Negotiables:
[Action]
[Action]
Quick Wins:
[Action]
[Action]
Watch Closely:
[Metric and what change to look for]
Flags for Medical Attention:
[Only include if genuinely warranted by the data]
TONE AND APPROACH
Speak like a knowledgeable friend who has spent time studying this person's data. Direct, specific, and genuinely useful. Every observation must reference actual numbers from the uploaded files. Completely avoid generic health advice. If the data does not support a conclusion, say so clearly rather than speculating.
If the dataset covers fewer than 14 days, note that patterns are suggestive rather than statistically reliable.
Avoid adding medical disclaimers unless something genuinely warrants clinical attention. This serves strictly as a performance and lifestyle analysis.What comes back
Claude works through the prompt in five stages. Here is what each one surfaces from your Whoop data.
- Stage 1. File inventory: Reads your CSV, identifies every data type Whoop captured, flags the date range and any gaps, and notes Whoop-specific context before doing anything else.
- Stage 2. Personal baseline: Calculates your actual averages for recovery score, HRV, strain, sleep performance, respiratory rate, and sleep stages. It relies entirely on your personal averages.
- Stage 3. Pattern analysis: Finds what is actually driving your recovery score up or down. HRV trends over 28 days. Strain versus recovery balance. What alcohol does to your specific HRV. Whether your bedtime variance is doing more damage than your total sleep hours.
- Stage 4. Top 5 leverage points: Ranks the five physiological issues doing the most damage to your metrics, with the specific Whoop data behind each one and concrete actions to fix them this week.
- Stage 5. Weekly protocol: Non-negotiables, quick wins, the exact data points to watch closely in the Whoop app, and anything that warrants a doctor's attention.
What it actually finds
These are the patterns Whoop never shows you in the app, but that are sitting in your raw export right now.
- Why your recovery score was low on specific days: What happened the night before and the day before that. Hard training session, late meal, alcohol, high stress. Correlated across your full history.
- Whether your bedtime is doing more damage than your sleep duration: Whoop tracks when you actually fall asleep. Bedtime variance above 45 minutes is often a bigger driver of poor recovery than total sleep hours. The prompt calculates your variance specifically and tells you what it is costing you.
- What alcohol is doing to your HRV specifically: Average HRV suppression from alcohol is 10 to 20% but your number is entirely unique. If you use the Whoop journal to log alcohol, the prompt calculates your personal suppression from your own data.
- Whether you are training hard on days your body is already under stress: Whoop's strain and recovery data combined shows exactly how often you are pushing hard on days your body has flagged as low readiness. The prompt quantifies this across your entire dataset.
- Temperature changes before symptoms appear: Whoop tracks skin temperature. Deviations from your baseline often precede illness by one to two days. The prompt shows you every instance where this happened in your history and what followed.
If something goes wrong
- You did not receive the export email: Check your spam folder first. If it is not there, go back to Whoop app, More tab, App Settings, Data Export and try again. Make sure the email address on your Whoop account is correct.
- Claude says the file is too large: If you have several years of Whoop data the CSV can be large. Try exporting a shorter date range from within the app. 90 days is enough for strong pattern analysis.
- Claude cannot read the file: Make sure you are uploading the CSV file directly. Avoid using a ZIP or a screenshot. The Whoop export should arrive as a CSV attached to the email.
- The analysis feels generic: This usually means the export did not include journal data. If you use the Whoop journal to log alcohol, caffeine, stress, or other behaviours, make sure those are included in your export settings before exporting.
How to actually use this
Three ways to get more from it.
- Use at least 90 days of data. Two weeks surfaces patterns. Three months surfaces the ones that actually matter. Cumulative fatigue, seasonal trends, and the real impact of your training cycles all need time to become visible.
- Log consistently in the Whoop journal. The journal is where you note alcohol, caffeine, stress, travel, and illness. The more behavioural data in the export, the more specific the output. If you have not been using the journal, start now and re-run this prompt in 30 days.
- Run it again after making changes. Export new data after four weeks of acting on the weekly protocol and run the same prompt again. The before and after comparison is where the real value shows up.
Where I'm taking this next
Most people check their Whoop recovery score, feel good or bad about it, and move on without knowing what caused it or what to actually change. The score becomes background noise. This prompt makes the data mean something. Your specific HRV response to alcohol. Your specific bedtime variance and what it costs your recovery score. Your specific strain-to-recovery ratio and whether it is trending in the right direction.
Whoop has been collecting this data every night. Now you get to read it.
