bgl-read

Note

This is based on the Sensotrend/sensotrend-uploader.

I just needed to be able to read data from a Contour Next One (older USB model) to grab the BGL data for someone else.

Claude did all the hard yards figuring out how to convert the small portion I needed into Rust.

Overview

Reads blood glucose readings from a Bayer/Ascensia Contour Next USB meter and outputs them as JSON or CSV.

Tested against the Contour Next One (USB model). Other Contour Next variants that use the same ASTM-over-HID protocol should also work — see docs/devices/contour-next.md for protocol details.

---

Building

cargo build --release

The binary ends up at target/release/bgl-read.

On macOS you may need to grant input-device access the first time you run it; the OS will prompt you.

---

Running

Plug the meter in via USB, then:

# JSON (default) — device info + all readings
bgl-read

# CSV
bgl-read --format csv

# Show a live progress line while reading
bgl-read --progress

# Write to a file instead of stdout
bgl-read --format csv --output readings.csv

# List connected Contour devices
bgl-read --list

# Capture the raw packet stream, then convert it without the meter
bgl-read --format binary --progress --output session.bin
bgl-read --from-bytes session.bin --format csv --output readings.csv

Reading the meter only once

Every run that talks to the meter is a full transfer of up to 800 records, so producing four formats used to mean four transfers. --format binary writes the raw HID packet stream to a file, and --from-bytes replays that file through the same framing and parsing code the live read uses — so a single transfer can produce every format, the bytes hex dump included.

bgl-read --format binary --progress --output session.bin
bgl-read --from-bytes session.bin --format records --output session.txt
bgl-read --from-bytes session.bin --format csv     --output session.csv
bgl-read --from-bytes session.bin --format json    --output session.json
bgl-read --from-bytes session.bin --format bytes   --output session.hex

bin/capture-bgl <prefix> does exactly this and leaves the files in captures/<date>/.

--from-records FILE is the narrower version: it re-parses a saved records dump, which is enough for csv and json but cannot reproduce bytes or binary, because the text dump does not carry the HID traffic. Both replay flags are offline-only, so neither accepts --progress.

Damaged input is reported rather than quietly producing a short reading list: --from-bytes rejects a truncated or foreign file outright, and --from-records warns on stderr with a count of the lines it could not parse.

Output formats

Formats are chosen with --format, and --output FILE writes to a file instead of stdout. binary refuses to write to a terminal — give it --output or pipe it somewhere.

--format	Description
json	Device info and readings as pretty-printed JSON
csv	One reading per row, header included
records	Raw ASTM record text as received from the meter (useful for debugging)
bytes	Hex dump of every HID packet exchanged (TX and RX)
binary	Compact binary packet capture, replayable with --from-bytes

Timestamps are in the meter's own local time (no timezone attached — the device has no concept of timezone).

---

Supported devices

All Contour Next devices share USB vendor ID 0x1A79 (Ascensia / Bayer).

Product ID	Device
0x7800	Contour Next One (USB-only variant)
0x7440	Contour Next USB
0x7350	Contour Next
0x7900	Ascensia Contour Next
0x6220	Contour Next Link
0x6230	Contour Next Link 2.4

The tool opens the first matching device it finds. Run --list to see what is connected.

---

Notes

• The meter stores the 800 most recent readings in a circular buffer — older readings are silently overwritten. Read regularly if you need a complete history.
• Control solution readings (used to verify the meter) are detected via the annotation field and excluded from output.
• --format records output is the raw meter transcript: the header (H) line includes the meter's password and serial number, and the dump contains every reading. The bytes and binary captures hold that same transcript verbatim. Redact these before sharing dumps publicly (e.g. when filing issues).