Train Commute Companion
Get an alert when delays, cancellations, platform changes or missed connections affect your usual train journey.
Choose a fixed train or any train within a time window.
Data: NS. You bring your own NS API key; none is bundled. This project is not made by or linked to NS.
What it does
- Follows your trains. You describe your usual trips once, in a small YAML file, with station names such as
Utrecht Centraalor NS codes such asUT. On a travel day it picks your train and watches it. - Tells you what changed. Delays, cancellations, platform changes, a train cut short, a new route, a transfer at risk or broken, and a train you missed.
- Suggests what to do. When your train no longer works, it names the best train you can still catch.
- Shows when to leave. Check-in and pre-trip advice use your travel buffer; a delayed train never moves leave-by later. Settings.
- Sends a morning commute brief. See today's trips and when to leave. The evening check alerts you if tomorrow's usual train needs attention. Daily brief settings.
- Looks ahead. It warns you about planned works on your route. Choose how many calendar days or upcoming commute dates to check. Timetable lookahead.
- Summarizes the week. An optional works digest checks seven calendar days and labels unknown coverage.
- Shows how your commute went. See your on-time rate and delays in local commute statistics.
- Tracks replacement transport. Follow reviewed bus, walk and metro plans with explicit boarding and arrival confirmations. Replacement transport.
- Shows crowd forecasts. NS forecasts appear with your trains and alternatives; changed forecasts can send silent updates. Crowd forecasts.
- Changes plans and alerts. Resume a skipped journey and control quiet or paused notifications in Home Assistant.
- Checks any station pair. See up to three upcoming train options in Telegram or Home Assistant.
- Keeps regular commutes and flexible trips together. A connected MCP agent can list both and preview departures for a saved trip. Trip changes can ask for approval in Telegram. MCP setup.
- Pauses for holidays. Set an inclusive date range and monitoring resumes afterward. Holiday pauses.
- Uses weekday departure times. Set different first trains within one profile while preserving already-followed journeys. Schedules.
- Follows any train in a time window. No usual train needed, with an optional minimum stay and settings per day. Schedules.
- Plans around when you need to arrive. Set an arrival target, include your walk, and choose how much earlier you would leave for a direct train. Arrive by.
- Notices another boarding station. With location on, it follows a known station by itself and asks once at a new one. Other boarding station.
- Shows train context. Optional NS composition, listed capacity and candidate vehicle coordinates include source limits. Train context.
- Shows commute weather. Station forecasts and tips need no key. Weather can suggest extra minutes in leave-by text. Official KNMI warnings need your own key. Weather.
- Loud only when it matters. Alerts you must not miss ring; small updates arrive silently.
- Says when NS is unclear. When NS gives two different answers about the same train, you get both, so you can decide.
Ways to run it
Compare setup options for Home Assistant, Telegram, and ntfy.
| Way | What you get | Guide |
|---|---|---|
| Home Assistant (HACS) | Alerts on your phone through the companion app, sensors and buttons | docs/home-assistant.md |
| Docker | Alerts and guided travel plans in a private Telegram chat (amd64 and arm64) | docs/docker.md |
| Docker with Home Assistant | One Docker schedule with Home Assistant sensors, commute controls and phone location for connected profiles | Shared schedule setup |
| ntfy | One-way phone push without Home Assistant or Telegram; your own private topic and access token | docs/ntfy.md |
| Household | A separate Docker companion and private bot for each traveller | docs/household.md |
| Webhooks | Each alert as signed JSON to your own URL, from Docker or Home Assistant | docs/webhooks.md |
| Optional encrypted SMTP copies alongside ntfy or Telegram in Docker | docs/email.md | |
| Today | Private journey and service status with browser refresh, a home screen shortcut and optional sent-alert history | docs/docker.md |
| MCP | Journey status and optional plan management for local or private-network agents | docs/mcp.md |
Location (optional) helps it know when you leave and where you are. See Where you are below.
Profiles can request accessible journeys and disable notification categories. These settings also appear in the Home Assistant profile form.
Save an explicit arrival to compare it with the original plan and optionally check NS compensation rules. The companion does not decide eligibility. Today also offers deliberate calendar and offline route-preview downloads.
Docker can keep one silent journey status pin up to date in your confirmed private Telegram chat.
With reminders enabled, confirm an NS-reported arrival with one tap. Before travel, open your boarding station in Maps.
Quick start (Home Assistant)
- In HACS, add
https://github.com/rhnfzl/train-commute-companionas a custom repository of type Integration, or use the button above. - Download Train Commute Companion and restart Home Assistant.
- Add the integration under Settings, Devices and services, and enter your NS API key.
- Choose stations and trains in the forms that follow, or turn on Use an existing YAML file to use a profiles file you already have.
The full guide is in docs/home-assistant.md.
Where you are
A timetable app knows your train. It does not know where you are. With location on, the companion can:
- see that you are still at home when your train leaves, and tell you the next train and when to leave;
- see that you are on the train, so a late phone position does not cause a false "you missed it";
- pick the next train you can catch when you leave home before your usual train leaves.
- notice that you are at another station and follow a train from there, or ask you first.
Phones often report a position minutes late. A position just after the departure does not prove a miss. The companion also checks movement, Wi-Fi, activity and time:
Activity counts only when reported after the departure and at most 5 minutes ago. Without proof the train stays followed, and you may get a quiet "has left. If you are not on it, next train" note. Regular train alerts still apply. Full rules: Missed trains and Other boarding station. Other sources: OwnTracks, Telegram.
Phone setup (Home Assistant companion app)
- Location: Always. iOS: Settings, Privacy & Security, Location Services, Home Assistant: Always, Precise Location on. Android: Allow all the time.
- Motion (optional, recommended). iOS: the app does not ask by itself. In the app open Settings, Companion App, Sensors, tap the blue Motion link and allow access, then enable Activity. Android: enable Detected activity. Details.
- Check it. In Home Assistant, Developer tools, States: your phone tracker has
latitudeandlongitude, and the activity sensor is notunavailable.
Then select your person and zones in the profile editor. The same settings are available in YAML. Location with Home Assistant.
The Python core
The core is a typed Python library on PyPI. It has no Home Assistant code.
It is released for its two shells, the Home Assistant integration and the Docker app, so its module APIs can change between versions. The supported entry point is the tcc command: check-profile, replay and run.
pip install train-commute-companion
tcc check-profile profiles.yaml
Development
uv run --extra app pytest -q # core and app tests (the extra installs the MCP SDK)
uv run --extra app mypy src app/tcc_app
# the published floors: Python 3.12, httpx 0.27, PyYAML 6.0.1 and mcp 2.3
uv run --no-cache --isolated --no-project --python 3.12 --with '.[app]' --with httpx==0.27.0 --with PyYAML==6.0.1 --with mcp==2.3.0 --with pytest --with pytest-asyncio pytest -q tests tests_app
uv run ruff check src tests tests_ha tests_app tests_live custom_components app
# Home Assistant integration: Python 3.14.2+, in its own environment
export UV_PROJECT_ENVIRONMENT=.venv-ha
uv run --python 3.14 --group ha pytest tests_ha -q
uv run --python 3.14 --group ha mypy --explicit-package-bases --namespace-packages -p custom_components.train_commute_companion
The Validate workflow runs these same commands, and a release runs them before it publishes. Each night it also runs tests_ha against the newest Home Assistant test package, because the integration calls private mobile_app code.
Live NS checks run separately with your own key and capture alerts locally.
License
Metadata
Release files for train-commute-companion 0.8.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| train_commute_companion-0.8.0.tar.gz | 1.7 MB | With provenance | Details for train_commute_companion-0.8.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| train_commute_companion-0.8.0-py3-none-any.whl | Python 3 | none | any | Details for train_commute_companion-0.8.0-py3-none-any.whl |
Total release size: 2.1 MB
Release files / train_commute_companion-0.8.0.tar.gz
| Download URL | train_commute_companion-0.8.0.tar.gz |
|---|---|
| Size | 1.7 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
3f169cce35e87453e8b12cff9ee177e697d94065c3fbf939cc6c0c4246a4ea6d
|
|
BLAKE2b-256 checksum How to use checksums |
4468645bc3d3c042557d69286f0a8a5f7b717e2bbccda098fffc226797be8698
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Nov 21, 2026.
Transparency logRelease files / train_commute_companion-0.8.0-py3-none-any.whl
| Download URL | train_commute_companion-0.8.0-py3-none-any.whl |
|---|---|
| Size | 391.7 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
aa14419342420c84dd48cc081d01fda0b3d1879ed4a7151606c6295a3edbea2e
|
|
BLAKE2b-256 checksum How to use checksums |
cc9da11a72ed4e54ee288d9a18224f14dcb97f5c07be278cb6b6bc82ffc6babd
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Nov 21, 2026.
Transparency logSecurity for train-commute-companion 0.8.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
Both files in this release have been uploaded with provenance metadata.
Attestation for train_commute_companion-0.8.0.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Nov 21, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.