How to install and set up Train at Home on macOS (no extra app needed)

Installing Train at Home itself doesn't require any third-party app — it's a single official installer. This guide covers the install and the first few minutes after launch; for what to do to keep it running well afterward, see the optimal setup guide.

Where to get it

Per the official user guide, the install path is: visit iota.macrocosmos.ai↗, click Download in the Train at Home section, then double-click the downloaded .dmg file to install the app. Don't download it from anywhere else — a .dmg from a random link isn't something you can verify.

What macOS supports today

The official user guide states plainly: "Currently TAH only supports MacOS, with Linux support coming soon." If you're on a Mac, you're in the supported case. This guide and subnera were last verified against macOS 26.6.2 and Train at Home 3.7.0.

Installing the .dmg

1. Open the downloaded .dmg file.
2. Follow the installer's own prompts to place the app where it belongs.
3. Launch it, and grant any system permission prompts macOS shows the first time (network access, in particular, is expected — the app needs it to register and train).

What "Connected" means

After the first launch, wait for the app's own status indicator to show Connected. That's the app's signal that it has reached the network and is ready to proceed — don't start troubleshooting anything before you see it.

What happens right after that

Once connected, the app registers with the orchestrator and moves through a short sequence of states before it starts doing real work: something like queued (waiting for a slot), then registering — where the app shows confirmed once registration itself is accepted, which only means it's still waiting for a layer slot, not that one has been assigned; this step can last hours — and only once a layer slot is actually assigned does it move to working (actually training a layer). The official app's own interface doesn't reliably surface exactly where you are in that sequence — that's a known limitation, not something wrong with your install.

Two ways to see it anyway:

• Read the CLI log yourself — this guide shows exactly what to look for.
• Install subnera, a small read-only menu bar app built specifically to surface this — see how to install it.

Neither is required to use Train at Home; they only make an already-running install easier to watch.

Right after install: the two things worth doing immediately

1. Set up a keep-awake method. Per the official FAQ, "Training stops automatically. Nothing continues running when the app is closed or your computer goes to sleep." A fresh install that goes to sleep overnight simply stops — see the keep-awake guide for exact commands.
2. Leave the Mac on a stable network and on power. The official guidance is: "Keep the laptop powered and on a stable network for steady rewards." No specific speed is required — see the optimal setup guide for what "stable" means in practice.

Where Train at Home keeps its files

Two locations are worth knowing about, even though a normal install doesn't require touching either:

• Logs: ~/Library/Logs/IOTA Train at Home/<YYYY-MM-DD>-cli.log, a new file each day. This is what every log-reading guide on this site — and subnera — actually reads. You never need to edit it; reading it, or letting a tool read it, is enough.
• The app itself, wherever the installer placed it — typically your Applications folder, like any other macOS app.

Knowing the log's location is useful the first time you want to check something yourself, without installing subnera or enrolling a subnera hub machine: running tail -f on today's file in Terminal shows you new lines as the app writes them, which is a reasonable way to confirm it's actually doing something even before you trust any higher-level tool.

If the install or first connection doesn't work

A few things worth checking before assuming something is broken:

• Confirm you downloaded from the official domain and that the .dmg fully finished downloading (a truncated download is a common, boring cause of install failures for any app).
• Confirm your Mac has a working internet connection at all — the app needs to reach the network before it can show Connected.
• Check that macOS's own network-access permission prompt wasn't dismissed by mistake on first launch; the app needs outbound access to reach the orchestrator.
• If the app opens but seems stuck before Connected, give it a little time; a first connection can take longer than a subsequent one.
• Restarting the app before you've ever registered isn't the same as restarting once you're already queued — with no queue position yet, there's nothing to lose by trying again at this early stage.

If none of that resolves it, the troubleshooting guide covers the more persistent failure modes (stuck queue, no routable peers, unexpected stops) once you're past the install step itself.

Quick checklist

1. Download only from iota.macrocosmos.ai↗.
2. Confirm you're on macOS — Linux isn't supported yet.
3. Open the .dmg, install, launch.
4. Wait for Connected before doing anything else.
5. Set up keep-awake right away (guide).
6. Optionally install subnera to watch the real status (guide).
7. Read the full setup guide for everything else worth knowing.