The Home Dashboard: Building the Start Screen

The Home Dashboard: Building the Start Screen

Module 13 · Lesson 8

Every single lesson so far in this module has been preparation, card types, names and icons, charts. This lesson finally builds a real, working view, the Start screen, the very first thing any household member sees when they open Home Assistant, and it needs to earn that position honestly.

What genuinely earns a place on Start

Start genuinely holds only what a household member needs across the entire house at a glance, whether any light is left on, today's weather, the movie-night scene, and a handful of the single most-used controls. Nothing here is genuinely room-specific detail, that kind of detail genuinely belongs on the Office view this module builds next, Start's whole job is a fast, whole-house overview, not an exhaustive, all-encompassing control panel for the entire house.

What genuinely doesn't belong on Start

Diagnostic entities, test_mode_hallway, timer countdowns, ESPHome uptime sensors, genuinely have absolutely no place here at all, they belong on the dedicated diagnostic view this module rebuilds a few lessons from now. Neither does every single entity in the house, Start is deliberately curated, and cramming every light, sensor, and helper onto it defeats the entire purpose of having a fast overview screen in the first place.

Start doesn't replace any other view

It's genuinely tempting to try making Start do everything, one screen for the whole house, but that instinct is exactly what this module's opening lessons warned against. Start, Office, Diagnostic, and Admin are four genuinely separate views with four separate audiences, and Start's value comes precisely from staying narrow, a jack-of-all-views Start screen quickly turns back into the exact cluttered mess this module exists to fix.

Layout: top, middle, bottom

A genuinely simple three-band layout works well for Start, a top section for status, weather, any active alert, a middle section for the handful of whole-house controls most used daily, and a bottom section for quick, genuinely simple links into Office and any other room views that exist. This top-to-bottom order genuinely mirrors how someone's eye naturally scans a screen, most urgent information first, action second, navigation last of all.

An eight-card limit, and why it genuinely matters

This course genuinely recommends capping Start at eight cards total, a number chosen deliberately, not arbitrarily, eight cards fit comfortably on one phone screen without scrolling, and forcing yourself to stay within that limit means every single addition has to justify bumping something else off. That constraint alone does more to keep Start genuinely useful than any amount of careful card-type selection ever could on its own.

Step by step: building Start in the UI

Create a new dashboard view named Start, using the Sections view type established earlier in this module. Add a status section at the very top with a tile for any active alert and a weather forecast card, a controls section in the middle with tiles for the hallway and office LEDs plus a button for scene.movie_night, and a navigation section at the bottom genuinely linking through to the Office view. Save, then genuinely count your cards carefully, if you're over eight, cut something before adding anything else at all to the view.

A YAML example: the Start view

title: Start
type: sections
sections:
  - type: grid
    cards:
      - type: tile
        entity: switch.led_status_hallway
        name: Hallway LED
      - type: tile
        entity: switch.led_status_biuro
        name: Office LED
      - type: button
        entity: scene.movie_night
        name: Movie Night

Notice how sparse this genuinely is, three cards, each one earning its place, this is precisely the discipline an eight-card limit is genuinely meant to enforce right from the very start.

What can genuinely go wrong on Start

Start can genuinely creep back toward clutter within weeks if every new automation's test toggle gets casually added "just for now." A diagnostic entity left on Start by mistake confuses a household member who has no context for what test_mode_hallway even means. And an eight-card view that's secretly nine or ten because one single card was forgotten during a careful count genuinely undermines the entire discipline this whole lesson is built around.

An exercise: build Start with a maximum of eight cards

Build your own Start view right now, genuinely capped at eight cards, using the top/middle/bottom layout from this lesson. Before adding a ninth card, genuinely ask yourself what you'd actually remove to make room for it, if the honest answer is "nothing," that's the clearest possible sign the entity you're about to add doesn't genuinely belong on Start at all.

A real-life example: a guest glancing at Start

Picture a guest staying over for the weekend, opening Home Assistant for the very first time, with genuinely no context for any automation, helper, or diagnostic concept covered across this entire course. A well-built Start view should be immediately understandable to that guest within seconds, a couple of lights, the weather, one movie-night button, nothing requiring an explanation. If Start would genuinely confuse a first-time guest, that's a strong signal it has drifted away from its actual intended purpose.

Weather as the one genuinely universal Start card

A weather forecast card genuinely earns a permanent spot near the top of Start in almost every household, it's information nobody has to remember to check separately, and it directly informs decisions elsewhere in the house, whether to close blinds, whether to expect the heating to work harder later. Unlike most other Start candidates, weather doesn't compete for the eight-card limit on entity-by-entity merit, it's simply default, universally relevant information.

Navigation cards: pointing to Office without duplicating it

The bottom navigation section genuinely shouldn't try to summarize the Office view's contents, a temperature reading here, a scene button there, that's exactly the duplication this module has warned against since its very first lesson. A single, clearly labeled link or button pointing at the Office view is genuinely enough, letting Office itself carry all of its own room-specific detail without Start trying to preview it.

Revisiting Start after a week of genuine daily use

A first attempt at Start is genuinely a hypothesis about what the household actually needs at a glance, not a final answer, and after a week of real daily use it's worth honestly asking which cards actually got tapped and which were quietly ignored the entire time. Swapping an unused card for something genuinely more relevant, based on real behavior rather than a guess made before the view even existed, is exactly the kind of iteration this course has encouraged since Module 12.

Why Start is genuinely the hardest view in this module

Counterintuitively, Start, the simplest-looking view in the entire module, is genuinely the hardest one to get right, precisely because its constraint is so tight. Office, Admin, and Diagnostic can each grow to fit whatever they genuinely need to show, Start alone has to survive an eight-card limit while still representing the entire house, and that discipline is exactly what makes it worth this lesson's full, careful attention rather than a quick afterthought.

Alerts: the one exception genuinely worth breaking the layout for

A conditional card, introduced back in the card-types lesson, genuinely belongs at the very top of Start, showing only when something genuinely needs attention, a door left open, a battery critically low, and staying invisible the rest of the time. This is the one deliberate exception to the eight-card discipline worth making room for, an alert that only appears when it's genuinely relevant costs nothing the other 99 percent of the time it stays hidden.

Choosing which scene genuinely belongs on Start

A household is likely to accumulate several scenes over time, movie night, a morning routine, a guest-arrival preset, and Start genuinely has room for only one or two of the very most frequently used ones, not an exhaustive scene library. Choosing based on actual, genuine daily frequency rather than personal favorites keeps Start firmly aligned with its real purpose, a fast, honest overview of what actually gets used every day, not a showcase of everything the system is technically capable of doing.

Whole-house tiles versus room-specific tiles

A genuinely useful filter for any candidate Start card is asking whether it represents something whole-house in scope, "are any lights on anywhere," rather than something tied to one specific room, "is the hallway LED on." The hallway and office LED tiles used in this lesson's YAML example are a deliberate simplification for a two-room course example, in a larger real home, a single aggregated "any light on" tile would genuinely replace several individual per-room tiles to stay within the eight-card limit.

Start as the honest measure of a well-organized house

A genuinely well-organized smart home shows itself precisely in how calm its Start view can afford to be, a house with dozens of unresolved automation quirks tends to accumulate warning tiles and test toggles on its main screen out of necessity, while a house with properly layered helpers, scripts, and dashboards, exactly what this entire course has genuinely built module by careful module, can keep Start down to a genuinely small, calm handful of cards worth actually looking at. Start's simplicity is a symptom of good underlying structure, not a decorative choice made independently of everything else.

Naming the Start view itself clearly

Just as Lesson 6 covered naming entities carefully, the view itself genuinely deserves a clear, unambiguous title, "Start" or "Home" rather than a default auto-generated label, so that anyone switching between views on a tablet or phone can tell at a glance which one they're currently looking at. This small naming detail matters considerably more once Admin, Diagnostic, and Office all exist side by side in the same view-switcher.

Testing Start with someone outside the project

The single most genuinely useful test for a finished Start view is handing a phone or tablet to a household member who had no part in building it, and simply watching what they do, do they immediately understand every card, do they hesitate over any label, do they tap something expecting a different result. Their genuinely honest, unrehearsed reaction reveals considerably more about Start's actual real quality than any amount of your own careful, biased review of your own work ever realistically could.

Start on a wall-mounted tablet versus a phone

A household with a wall-mounted tablet running Home Assistant continuously might genuinely be tempted to loosen the eight-card limit there, since screen space is no longer scarce the way it is on a phone. This course still recommends keeping Start's actual card count consistent everywhere it appears, a wall tablet can certainly render those eight cards larger and more comfortably spaced, but discipline about what belongs on Start at all should genuinely hold regardless of which physical screen happens to be displaying it.

Seasonal or temporary changes to Start

It's genuinely fine, even sensible, to temporarily swap a Start card for something seasonal, a holiday-lighting scene button replacing movie-night for a few weeks in December, provided it goes back afterward rather than accumulating permanently. Treating Start as something that can flex briefly for genuinely relevant, temporary needs, while still respecting the eight-card ceiling at every point in time, keeps the view responsive to real life without ever letting it drift back into permanent clutter.

Connecting Start back to this module's opening three questions

Every card genuinely placed on Start in this lesson traces directly back to the three questions this module's very first lesson introduced, what's happening right now, the status section, what can I change right now, the controls section, and is anything wrong right now, the conditional alert card. Any candidate card that can't be mapped cleanly onto one of those three questions genuinely doesn't belong on Start, no matter how convenient it might otherwise seem to include.

Start as a foundation the rest of this module builds on

Every remaining view this module still has to build, Office, the rebuilt Diagnostic dashboard, Admin, and both the mobile and desktop layouts, genuinely assumes Start already exists as the calm, reliable entry point a household member returns to after visiting any of them. Getting Start genuinely right now, rather than rushing carelessly past it to reach the more visually interesting views still ahead, is exactly the kind of front-loaded discipline this entire module has been genuinely building toward since its very first lesson, and it pays off directly in every view still to come.

Key takeaways

Start holds only whole-house overview items, never room-specific detail or diagnostic entities.

An eight-card limit forces genuine curation, every addition must justify bumping something else.

A top/middle/bottom layout, status then controls then navigation, mirrors how eyes scan a screen.

Start doesn't replace Office, Admin, or Diagnostic, each view keeps its own narrow purpose.

What's next

With Start built and genuinely disciplined, the next lesson zooms into a single room, the Office, as the pattern every other room view in your house will follow.

Finished this lesson?