PulseCourier: Health Export
iPhone / Forme et santé
Your health data, ready for any AI in one tap.
PulseCourier exports everything recorded in your Health app into a single file you can paste into ChatGPT, Claude, Gemini, or feed to your own scripts. Pick a date range, tap once, share the file. That's it.
—
WHAT GETS EXPORTED
Steps, walking and running distance, active energy, exercise and stand minutes, heart rate, resting heart rate, HRV, sleep stages (in bed, core, deep, REM, awake), and every workout with duration, distance, calories, and per-workout heart rate.
—
PICK A RANGE
Today, last week, last month, or any custom date range. Or switch to Select Sessions mode and pick specific workouts and sleep nights one by one.
—
COPY WITH AI PROMPT
The "Copy with AI Prompt" button puts two things on your clipboard at once: a ready-to-use analysis prompt and the export file itself. Open ChatGPT, paste, hit send. Your data lands together with a prompt that tells the model what to do with it.
—
THREE FORMATS
NDJSON for AI tools and scripts. CSV for spreadsheets. Markdown for a human-readable summary with tables for activity, heart, sleep, and workouts.
Canonical 5-minute buckets by default, so simultaneous iPhone and Apple Watch samples don't double-count. Switch to raw samples for source-level dumps when you need them.
—
PRIVACY
No accounts. No servers. No tracking. PulseCourier has zero networking code. Your health data only leaves your iPhone when you share the file through the system share sheet. Health permissions are read-only.
—
WHAT IT'S NOT
PulseCourier is not a background syncer or a charts app. It does one thing: turn your health data into a clean file, fast. If you want automated daily exports to Dropbox or graphs inside the app, you want a different tool. If you want to hand your data to an AI in 5 seconds, this is it.
Quoi de neuf dans la dernière version ?
/Exports are now significantly smaller. Raw health metrics are now encoded as compact columnar arrays – same data, same precision, but around 65% smaller files and far fewer tokens.
