All guides

How to monitor an Augment (Auggie) session from your phone

A practical workflow to start Auggie on a computer you trust, then monitor progress, handle approvals, and send follow-ups from your phone without SSH.

Based on
CLI
0.1.0
Preview ref
f36aa45

I like Augment’s “just start coding” vibe.

What I don’t like is babysitting my laptop while an Auggie session runs.

If a task is going to take 20–60 minutes, I want a workflow where I can:

  • start the session on the computer that has the repo
  • put the computer down
  • check progress from my phone
  • answer any prompts/approvals
  • send one follow-up

This guide is that workflow.

TL;DR

  1. Install Auggie and verify auggie --version on the computer that will run the session.
  2. Use a phone workflow (Happier is one option) to monitor the session, handle approvals, and send follow-ups.
  3. If Augment auth status shows Unknown, don’t panic — Auggie auth probing is conservative.
  4. If Auggie isn’t found, set HAPPIER_AUGGIE_PATH to the full path of the auggie binary.

The problem

“Monitor from my phone” usually fails for one of three reasons:

  • the session is running on a computer that sleeps (so the session dies)
  • you only have SSH (so your phone becomes a terminal)
  • auth/status is ambiguous (so you don’t trust what you’re seeing)

Auggie is a local CLI-backed provider. That’s a feature — the code stays where your repo is.

But it means your phone workflow needs to be intentional.

What I tried first (and where I bounced off)

  • SSH from my phone: works, but I don’t want to read long output in a tiny terminal.

  • Remote desktop: heavy and fiddly.

  • Just wait until I’m back at a keyboard: safe, but slow.

  • Happier: it lets me open the Augment session from my phone, check progress, handle approvals, and nudge the run with one short follow-up. Trade-off: it still depends on the computer running Auggie being online.

If you’re new to Happier (what it is, and where to get it)

If you haven’t seen Happier before: it’s an open-source companion app (mobile/web/desktop) for coding sessions.

For this guide, the relevant idea is simple:

  • Auggie runs on the computer that has the repo
  • Happier is where you supervise the session from your phone (progress, approvals, one follow-up)

Get set up:

Step 1: Install Auggie on the computer that will run the session

From the provider docs:

npm install -g @augmentcode/auggie
auggie --version

If you use hstack, there’s also an installer:

hstack providers install auggie
auggie --version

(If you haven’t seen hstack before: it’s an optional helper in the Happier ecosystem that can install provider CLIs for you. You don’t need it — npm install -g ... is totally fine.)

The important part is not the install method.

It’s that auggie --version works on the computer that will run the session.

Step 2: Understand Auggie authentication (and why status may be “Unknown”)

Auggie supports launching its native login flow from provider settings.

But Auggie auth probing is conservative.

In the Happier agent auth probe config, Auggie uses an unknown parser. That means:

  • Happier won’t try to infer “logged in” vs “logged out” with a risky probe
  • the UI may show Unknown even when everything is fine

So the rule I use is:

  • if auggie login works on that computer, you’re good
  • treat “Unknown” as “not safely probeable”, not “broken”

A practical login checklist

On the computer that will run the session:

  1. Run auggie --version.
  2. Run auggie login (or launch it from provider settings).
  3. Start one tiny session and confirm you can send a follow-up.

Once that works once, you can stop thinking about auth until it breaks.

Step 3: Fix “Auggie not found” with HAPPIER_AUGGIE_PATH

If the session computer can run auggie, but the app can’t find it, the fix is usually PATH.

Happier supports a binary override env var:

  • HAPPIER_AUGGIE_PATH

Set it to the absolute path of the auggie executable.

This override is exercised in the CLI backend tests (so it’s not just a doc promise).

Step 4: Start a session, then use your phone as the monitor

The phone workflow is intentionally narrow.

If I had to describe it as one loop:

  1. open the session
  2. check whether it’s waiting on approval
  3. send one follow-up
  4. leave

What I do from my phone (and what I don’t)

I’m comfortable doing this from a phone:

  • approve/deny a permission request after reading it
  • send a short correction (“stop after this step”, “don’t touch X”, “summarize first”)
  • sanity-check what files changed

I avoid doing this from a phone:

  • any big refactor
  • anything that needs lots of copy/paste
  • anything where I can’t easily verify the path/command

Three copy/paste follow-ups that work well for Auggie

If you want the “I’m on a phone” set of follow-ups, these are the ones I reuse:

  1. Checkpoint

Stop after your next step and summarize:

  • what you changed
  • what you didn’t change
  • what you want to do next
  1. Constraint

Don’t touch migrations/auth. If you think you need to, stop and ask.

  1. Diff-first

Before you keep going, list the files you modified and why (one sentence each).

Step 5: Troubleshooting checklist

The session isn’t updating

  • confirm the computer is online and the session runtime is still running
  • check whether it’s waiting on approval
  • reload/reopen the session

If you closed the laptop lid and that laptop is the host, assume the session stopped.

I can’t tell if it’s stuck

Use the “stuck session” playbook:

  • waiting for approval vs offline vs busy

If you’re new to Happier

If you haven’t seen Happier before: it’s an open-source companion app for running sessions like Augment/Auggie on one computer and monitoring them from your phone.

The workflow is basically:

  • Auggie runs where the repo is
  • Happier is where you check progress, handle approvals, and send follow-ups

(And if you decide you don’t want another app, the fallback is always SSH — but most people don’t love SSH on a phone.)

Get set up:

augmentauggiemobileworkflows
Last updated: 2026-04-03