How to export Apple Health data to CSV

Export from the Health app (official, free)

This is the only export Apple documents. It bundles everything the Health app stores, with no choice of metric or date range.

1. Open the Health app on your iPhone.
2. Tap Summary, then tap your picture or initials in the top right corner.
3. Scroll to the bottom of your profile and tap Export All Health Data.

   

4. Tap Export on the Export Health Data alert. With years of data, preparing the file can take a few minutes.

   

5. The share sheet shows the file as export, a ZIP Archive. Tap Save to Files, the easiest way to open it on a computer later.

   

6. On the computer, unzip export.zip. The records are in apple_health_export/export.xml.

Convert export.xml to CSV with Python

export.xml stores each measurement as a Record line with type, unit, dates and value. The script below streams the file, picks one data type and writes a CSV with start, end, value, unit and source. It needs only Python 3, nothing to install.

1. Save the code below as to_csv.py next to the apple_health_export folder.
2. In Terminal, in that folder, run: python3 to_csv.py apple_health_export/export.xml HKQuantityTypeIdentifierStepCount
3. The CSV lands in the same folder, named after the data type, and the last printed line says how many records were written. Replace HKQuantityTypeIdentifierStepCount with the type you want, for example HKQuantityTypeIdentifierHeartRate for heart rate.

import csv, sys
import xml.etree.ElementTree as ET

src, want = sys.argv[1], sys.argv[2]  # export.xml HKQuantityTypeIdentifierStepCount
rows = 0
with open(want + ".csv", "w", newline="", encoding="utf-8-sig") as out:
    w = csv.writer(out)
    w.writerow(["startDate", "endDate", "value", "unit", "sourceName"])
    events = ET.iterparse(src, events=("start", "end"))
    _, root = next(events)
    for event, el in events:
        if event == "end" and el.tag == "Record":
            if el.get("type") == want:
                w.writerow([el.get("startDate"), el.get("endDate"), el.get("value"), el.get("unit"), el.get("sourceName")])
                rows += 1
            root.clear()
print(rows, "rows in", want + ".csv")

What our test showed

We built a sample export.xml in Apple's format with 1 million step and heart rate records (220 MB, illustrative values, not anyone's data). On an Apple Silicon Mac with Python 3.15, on 2026-11-13:

Time is about the same; memory is not. An estimate, not tested: at the same ratio, a real 1 GB export would need close to 4 GB of memory to load whole, enough to freeze an ordinary laptop. The streaming script stayed near 16 MB in the test.

The paths side by side

When you do not need CSV

If the goal is to ask something about your data ("did my resting heart rate drop this year?"), exporting and converting every time is a chore. telo reads Apple Health through the iPhone app and lets Claude, ChatGPT and other assistants query the history directly, with no file. Developers can read the same data through the personal REST API, in JSON, up to 500 rows per call. If you only want spreadsheets, telo also hands its copy back as CSV: one row per day and metric, per night and per workout, in your time zone (Account > Your data on the web).

curl 'https://telo.health/api/v1/query_daily?metrics=steps&from=2026-11-01&to=2026-11-08&granularity=day' \
  -H "Authorization: Bearer $TELO_API_KEY"

Practical questions