Getting started¶
A first run, start to finish: install the client, log in, pull your positions into a local cache and write a GPX track. It takes a few minutes. Afterwards, the how-to guides cover the everyday tasks and the reference lists every command.
Note
This is an unofficial client. trackiwi publishes no API; the tool uses the private one its own app uses, which can change without notice. It is read-only: the only request that changes anything on the trackiwi side is the logout that revokes your session.
1. Install¶
You need Python 3.11 or newer and nothing else: the package has no runtime
dependencies. Check with python3 --version. If your system Python is older (older macOS,
for one), install a newer one and call it explicitly, for example python3.13.
pipx install git+https://github.com/ckeller42/trackiwi-client
trackiwi --help
There is nothing to build, so you can also run straight from a clone with
python3 -m trackiwi --help. It behaves exactly like the installed trackiwi command.
2. Log in¶
trackiwi login
It asks for your trackiwi email and password. The password is used once and never stored.
What is stored is the session token, in ~/.config/trackiwi/config.json with mode
0600. That token grants live vehicle location, so read
the security notes once.
Check that it works:
trackiwi trackers
You get one tab-separated line per device: id and name.
3. Sync¶
trackiwi sync
The first run fetches your whole history into ~/.local/share/trackiwi/positions.db
(mode 0600). Every later run fetches only what is new. If a sync is interrupted, run it
again: it resumes from the highest position already stored.
4. Export¶
trackiwi export --format gpx --from 2026-09-16 --to 2026-09-25 -o route.gpx
Open route.gpx in any GPX viewer. --format is gpx, geojson or csv; the dates
are inclusive UTC calendar days. With several devices, each keeps its own track.
5. Clean up when you are done¶
trackiwi logout # revokes the session on the server, then deletes the local token
trackiwi purge --yes # deletes the position cache
Where next¶
Everyday tasks: sync and export, and ingest into InfluxDB and Grafana.
Look something up: CLI, units, export formats.
Why it is built this way: architecture.