Build a Private Home Assistant Brain

Run a local Home Assistant voice pipeline and prepare an Echo Dot safety brief.

Introduction

30 Second Summary

Voice assistants are convenient until every request depends on an external service. An older Echo Dot can have a useful second life when its control center runs inside your home.

In this project, you will build a Home Assistant OS control center that operates a virtual Workshop Light through Assist with local speech services ready for a future Echo satellite. You will also use Claude Chat to create a safety brief before any hardware changes.

What You'll Build

Your Mac will show a Home Assistant dashboard where typing turn on Workshop Light in Assist visibly switches the Workshop Light on.

By the end of this project, you'll have:

  • Demonstrate a working smart-home brain when a typed Assist command changes the virtual Workshop Light on the Home Assistant dashboard.
  • Show a local voice pipeline with Speech-to-Phrase handling speech-to-text while Piper handles text-to-speech.
  • Use a Claude-reviewed safety brief to decide when the D9N29T Echo conversion must stop.
  • Secret Mission: Add a custom private-workshop sentence that toggles the virtual light and returns a custom confirmation.

Are there any prerequisites?

You need a Mac with internet access, administrator access, capacity for at least 2048 MB of RAM plus 2 virtual CPUs for a virtual machine, and a free Claude account or access to Claude Chat.

The Echo stays untouched throughout this project, so no soldering or firmware writing happens here.

Before We Start

A safe Echo conversion starts with a local Home Assistant system that can process smart-home requests privately. The Echo stays untouched while you build its local brain and a written safety gate.

Launch Home Assistant OS

The future converted Echo needs a reachable local server before any hardware work begins. Building that destination first keeps the risky Echo work outside this project.

In this step, you will use Oracle VirtualBox 7.2.22 to run Home Assistant OS inside a virtual machine on your Mac. Home Assistant OS provides the Apps system that later local speech services require.

In this step, get ready to:
  • Identify your Mac architecture.
  • Install the matching VirtualBox package.
  • Launch the matching Home Assistant OS virtual machine.
Match the downloads to your Mac

Intel Macs need different host software from Apple silicon Macs. The same architecture choice determines which VDI image Home Assistant OS can boot.

  • Click the Apple menu in the top-left corner of your screen.
  • Select About This Mac.
  • Identify whether the Mac summary shows an Intel processor or Apple silicon.
  • Record your Mac architecture here: Intel or Apple silicon.

Your architecture controls both downloads. Choose the tab that matches Intel or Apple silicon.

Intel Mac

Apple silicon Mac

The extracted .vdi file is the virtual hard drive that already contains Home Assistant OS. VirtualBox still needs to be installed before it can start that drive.

  • Press Cmd+Space to open Spotlight.
  • Type VirtualBox into Spotlight.
  • Press Enter if VirtualBox appears in the results.
  • Use the VirtualBox app menu in the macOS menu bar to open its version information.

Choose the tab that matches what you find.

✔️ I see version 7.2.22

Your matching VirtualBox installation is ready. That removes the first compatibility hurdle.

  • Close the version information window.

ⓧ I see an older version

macOS may request your administrator password during the upgrade. This is the expected permission check for installing system software.

  • Quit the older VirtualBox application.
  • Open the downloaded VirtualBox 7.2.22 host package from your Downloads folder.
  • Continue through the installer using its default installation location.
  • Enter your Mac administrator password if macOS requests it.
  • Press Cmd+Space after the installation finishes.
  • Type VirtualBox into Spotlight.
  • Press Enter to launch VirtualBox.
  • Confirm that the version information now reports 7.2.22.

ⓧ VirtualBox is not installed

macOS may request your administrator password during installation. This is the expected permission check for installing system software.

  • Open the downloaded VirtualBox 7.2.22 host package from your Downloads folder.
  • Continue through the installer using its default installation location.
  • Enter your Mac administrator password if macOS requests it.
  • Press Cmd+Space after the installation finishes.
  • Type VirtualBox into Spotlight.
  • Press Enter to launch VirtualBox.
  • Confirm that the version information reports 7.2.22.

VirtualBox Won't Install?

  • Confirm that the downloaded host package matches Intel or Apple silicon.
  • Review any macOS security prompt that appears after opening the installer.
  • Allow the installer through that prompt before trying the package again.

Get help with my VirtualBox installation.

Configure the Home Assistant virtual machine

The virtual machine needs a guest type that matches your Mac. Its memory allocation must also leave enough room for Home Assistant OS to run.

  • Start creating a virtual machine in VirtualBox Basic Mode.
  • Enter Home Assistant as the virtual machine name.

The Basic Mode summary should now show Home Assistant as the machine name.

Choose the guest settings that match Intel or Apple silicon.

Intel Mac

  • Select Linux as the guest type.
  • Select Oracle Linux (64-bit) as the guest version.

Apple silicon Mac

  • Select Linux as the guest type.
  • Select Oracle Linux (ARM 64-bit) as the guest version.

On Intel, the guest summary should show Oracle Linux (64-bit). On Apple silicon, it should show Oracle Linux (ARM 64-bit).

  • Set the memory allocation to at least 2048 MB.
  • Set the processor allocation to 2 CPUs.

The resource summary should now show at least 2048 MB of memory. It should also show 2 CPUs.

  • Enable Use EFI.
  • Complete the virtual machine creation flow in Basic Mode.

You should now see a virtual machine named Home Assistant in the VirtualBox machine list.

The storage controller must match your Mac architecture. This lets VirtualBox read the extracted Home Assistant OS disk correctly.

  • Select the Home Assistant virtual machine in the machine list.
  • Open its storage settings.

Intel Mac

  • Attach the extracted Home Assistant OS VDI under Controller: SATA.

Apple silicon Mac

  • Attach the extracted Home Assistant OS VDI under Controller: VirtioSCSI.
  • Save the virtual machine storage settings.

The storage summary should now list the extracted .vdi beneath the controller for your Mac architecture.

Why Do These Settings Matter?

Home Assistant OS requires UEFI firmware to boot. The Use EFI setting gives the virtual machine that firmware.

Intel and Apple silicon use different virtual storage controllers. Matching the controller to the architecture gives the VM access to its Home Assistant OS disk.

Can't See the Virtual Disk?

  • Confirm that you extracted the downloaded archive before looking for the .vdi file.
  • Confirm that the storage controller matches Intel or Apple silicon.
  • Return to Basic Mode if the Home Assistant virtual machine does not appear in the machine list.

Get help with my Home Assistant VM settings.

Bridge the VM to your network

Bridged networking places the virtual machine on the same local network as your Mac. Your browser can then reach Home Assistant through its local address.

  • Select the Home Assistant virtual machine in VirtualBox.
  • Open its network settings.

The network panel should now show the adapter configuration for the selected virtual machine.

  • Set Attached to to Bridged Adapter.
  • Select the Mac's active Wi-Fi or Ethernet adapter.

The network summary should now show Bridged Adapter with your active Mac network adapter.

  • Save the virtual machine network settings.
  • Start the Home Assistant virtual machine.

The first boot can take several minutes while Home Assistant OS prepares its services. Quiet stretches in the VirtualBox console are normal during this first start.

  • Wait for the current boot activity in the VirtualBox console to settle.

Before you check the address, do you expect the new virtual machine to be reachable from the Mac?

  • Open your Mac browser from the Dock.
  • Enter http://homeassistant.local/ in the address bar.

You should see Preparing Home Assistant or the Home Assistant welcome screen. That's the local brain online and reachable from your Mac.

Can't Reach Home Assistant?

  • Keep the Home Assistant virtual machine running while you test the address.
  • Return to the virtual machine network settings.
  • Confirm that Bridged Adapter uses the Mac's active Wi-Fi or Ethernet adapter.
  • Try the fallback address http://homeassistant.local:8123/ in your Mac browser.

Get help reaching my Home Assistant VM.

Your Mac now has a running destination for the future Echo. The Echo itself remains untouched.

Your Local Brain Is Online

  • Home Assistant OS is installed from the matching official VDI.
  • The Home Assistant virtual machine is running.
  • The guest type matches the Mac architecture.
  • The VM has at least 2048 MB of memory.
  • The VM has 2 CPUs.
  • The VM uses EFI with the architecture-specific storage controller.
  • The VM uses Bridged Adapter with the Mac's active network adapter.
  • Your browser reaches Home Assistant at http://homeassistant.local/.
  • No soldering or firmware work has been attempted on the Echo Dot.

Your Home Assistant server is ready. Next, you will finish onboarding and give it a virtual light that Assist can control.

Create Your First Controllable Device

Your Home Assistant server is now reachable from your Mac. The next proof is a device you can control from its dashboard.

A visible Toggle helper gives Assist a safe target for its first command. This proves the Home Assistant brain can act on a request before local speech services or Echo hardware enter the picture.

In this step, get ready to:
  • Complete Home Assistant onboarding with a local owner account.
  • Create an Assist-exposed Toggle helper named Workshop Light.
  • Turn on Workshop Light through typed Assist.
Finish Home Assistant onboarding

The preparation screen can sit quietly for several minutes. That is Home Assistant finishing its first boot.

Credentials deserve a careful moment. Your password stays off screenshots and out of project submissions.

  • Return to the browser tab showing http://homeassistant.local/.
  • Leave the tab open if you still see Preparing Home Assistant.
  • Select Create my smart home when the welcome screen loads.
  • Enter your name in the local owner account form.
  • Choose a username for your local owner account.
  • Create a strong password for your local owner account.
  • Submit the owner account form.

Your owner account now gives you administrative access to this Home Assistant installation.

  • Set your home location using the onboarding controls.
  • Review the sharing preferences on the next screen.
  • Choose the sharing options that fit your preferences.
  • Select Finish to open the default dashboard.

You are signed in to your own smart-home dashboard. That first setup hurdle is behind you.

Create and expose the Workshop Light

A Toggle helper is a virtual control with an on state and an off state. It gives you a visible device without requiring physical smart-home hardware.

  • Select Settings in the Home Assistant sidebar.
  • Select Devices & services.
  • Open the Helpers tab.
  • Select Create helper.
  • Select Toggle as the helper type.
  • Enter Workshop Light as the helper name.
  • Submit the form to create the helper.

Your virtual Workshop Light now exists as a controllable on-or-off entity.

Assist only controls entities that you expose to it. Exposure makes the Workshop Light available to the conversation pipeline.

  • Return to Settings.
  • Select Voice assistants.
  • Open Expose.
  • Select Workshop Light so it is exposed to Assist.

Workshop Light is now available to typed Assist commands.

Can't find Workshop Light under Expose?

  • Return to the Helpers tab to confirm that Workshop Light appears in the list.
  • Refresh the browser if the helper exists but the exposure list has not updated.

Help me expose my Workshop Light helper to Assist.

Control the Workshop Light with Assist

The dashboard gives you a visible place to watch the helper change state. Add the helper there only if Home Assistant has not already displayed it.

  • Return to the default dashboard.

✔️ Workshop Light is visible

The dashboard already contains a visible Workshop Light control. Its current state should show as off.

ⓧ Workshop Light is missing

Adding a card makes the helper visible on the dashboard. If Home Assistant asks you to take control, automatic updates stop for this dashboard.

  • Select Edit dashboard in the top-right corner.
  • Open the three-dot menu if the dashboard control dialog appears.
  • Select Take control.
  • Select Add card.
  • Choose By entity.
  • Select Workshop Light from the entity list.
  • Check the Unassigned section if Workshop Light is absent from the main list.
  • Select Done to save the dashboard change.
  • Select Assist in the top-right corner of the dashboard.

Before you send the command, do you think the dashboard control will change without you clicking it?

  • Type turn on Workshop Light into Assist.
  • Submit the typed command.

You will see the Workshop Light control change to on. That visible change proves typed Assist can control the helper.

There it is. Your Home Assistant brain just handled its first visible control request.

Workshop Light didn't turn on?

  • Check that the typed command says turn on Workshop Light with the helper name spelled exactly.
  • Return to Expose to confirm that Workshop Light remains selected.
  • Refresh the dashboard if Assist confirms the command but the visible control has not updated.

Help me troubleshoot my typed Assist command.

Your first controllable device is working. Next, you will inspect what this system still needs before it can listen and speak locally.

Identify the Missing Voice Engines

Your Workshop Light now responds to typed Assist commands. Next, you will test whether the same assistant is ready to handle audio.

A future Echo needs Home Assistant to turn incoming speech into text. Home Assistant must also turn its reply into speech.

In this step, get ready to:
  • Review the three stages available in a new voice assistant.
  • Check whether the audio stages have local providers.
  • Record what the current setup can process.
Map the assistant stages

An Assist pipeline divides a request into separate stages. Each stage handles one part of the journey from a request to a response.

What are the three stages?

  • The Conversation agent interprets text. It chooses the home-control action.
  • The Speech-to-text stage turns incoming audio into text for the conversation agent.
  • The Text-to-speech stage turns the response into audio for a speaker.
  • Select Settings in the Home Assistant sidebar.
  • Select Voice assistants to view the current assistant configuration.

The Voice assistants page brings the three pipeline stages into one configuration. A new assistant reveals which providers are currently available for each stage.

Test the local audio readiness

The provider menus reveal whether this fresh Home Assistant installation can process audio locally. This check leaves the working typed setup unchanged.

Before you inspect the provider menus, do you think both local audio stages are ready to select?

  • Select Add assistant.
  • Review the choices under Conversation agent without changing the current selection.
  • Open the Speech-to-text menu to check whether Speech-to-Phrase is available.
  • Open the Text-to-speech menu to check whether Piper is available.
  • Keep every provider choice unchanged.

You will see that the existing conversation path is available. That explains why typed Assist can control the Workshop Light.

Speech-to-Phrase is absent from the Speech-to-text choices. Piper is absent from the Text-to-speech choices.

That missing pair is the designed obstacle. The current setup can process typed requests without handling future Echo audio.

Can't see all three fields?

  • Confirm that you selected Add assistant from the Voice assistants page.
  • Scroll through the new assistant form to find each provider field.
  • Refresh the Home Assistant page if the provider menus do not load.

Need help locating the provider fields?

  • Use your browser's back button to leave the draft assistant without saving it.
  • Confirm that the Voice assistants page shows no replacement assistant from this check.

No replacement assistant should appear because you left every provider choice unchanged. Your Workshop Light remains on from the earlier typed command.

Why does typed Assist still work?

Typed Assist begins at the conversation agent. That is why the Workshop Light responded without a speech engine.

Future Echo audio needs a local speech-to-text provider. A local spoken reply needs a local text-to-speech provider.

You have isolated the exact boundary of the current setup. Next, you will add the two local engines that close it.

Build the Local Assist Pipeline

Your typed Assist test proved that the Home Assistant conversation agent can control Workshop Light. A future Echo still has no local way to pass audio in or receive speech back.

Speech-to-Phrase supplies focused local speech-to-text. Piper supplies local text-to-speech.

You will connect both services through Wyoming. Then you will test a local Assist pipeline against Workshop Light.

In this step, get ready to:
  • Run Speech-to-Phrase and Piper as Home Assistant OS apps.
  • Add both local voice services through Wyoming.
  • Build a Private Echo pipeline that turns off Workshop Light in Debug.
Install and start the local voice apps

Home Assistant OS apps package the services that process speech inside your virtual machine. These apps are available because your virtual machine runs Home Assistant OS.

  • Switch back to the Home Assistant dashboard from earlier.
  • Select Settings in the left sidebar.
  • Select Apps.
  • Select Install app.
  • Select Speech-to-Phrase from the app catalog.
  • On the Info tab, select Install.
  • When the installation finishes, select Start.
  • Leave Start on boot enabled.

You should see that Speech-to-Phrase is active. Home Assistant now has a focused local transcription service.

Why Speech-to-Phrase Instead of Whisper?

Speech-to-Phrase is close-ended. It supports a focused subset of Assist commands with lower processing needs.

Whisper is another local option for broader transcription. Speech-to-Phrase fits this project because the goal is focused home control on unconfirmed Mac hardware.

  • Return to Apps.
  • Select Install app.
  • Select Piper from the app catalog.
  • On the Info tab, select Install.
  • When the installation finishes, select Start.
  • Leave Start on boot enabled.

That is the voice pair in place. Home Assistant now has local services for both sides of a spoken exchange.

Voice App Not Starting?

  • Confirm that the running Home Assistant virtual machine still uses its bridged network connection.
  • Return to the app's Info tab to confirm that installation completed.
  • Select Start again if the service stopped after installation.

Help me troubleshoot a Home Assistant voice app that will not start. You can also search the Home Assistant Community.

Add the services through Wyoming

Each installed voice app runs as a background service. Home Assistant needs a Wyoming integration entry before that service can appear as an Assist provider.

  • Return to Settings.
  • Select Devices & services.
  • Find the Discovered section.
  • Select Add for the Speech-to-Phrase service.

You should see Speech-to-Phrase added as a local service. It is now available to supply speech-to-text.

  • Select Add for the Piper service.

You should see Piper added as a local service. It is now available to supply text-to-speech.

What Does Wyoming Do?

The Wyoming integration connects each running voice service to Home Assistant. This connection makes the service available inside an Assist pipeline.

Both services continue processing inside your Home Assistant machine. Spoken commands can stay in your home when a future audio device uses this pipeline.

Services Missing From Discovered?

  • Return to Apps to confirm that the missing service is started.
  • Reload the Home Assistant page after the service starts.
  • Return to Devices & services to check the Discovered section again.

Help me find a local voice service that Wyoming has not discovered. You can also ask the Home Assistant Community.

Configure and test Private Echo

The two provider entries are ready to become one local pipeline. The pipeline sends recognized text to the Home Assistant conversation agent before Piper prepares the spoken reply.

  • Return to Settings.
  • Select Voice assistants.
  • Select Add assistant.

The assistant form joins the conversation agent with one speech-to-text provider and one text-to-speech provider. Each selection controls one stage of the local exchange.

  • Enter Private Echo as the assistant name.
  • Choose your language.
  • Select Home Assistant as the Conversation agent.
  • Select Speech-to-Phrase as Speech-to-text.
  • Select Piper as Text-to-speech.
  • Select Save.

You should see Private Echo in the voice assistants list. Your local providers now belong to one saved pipeline.

How the Pipeline Fits Together

  • Speech-to-Phrase turns a focused spoken command into text.
  • Home Assistant interprets that text as a smart-home request.
  • Piper turns the response into local speech.
  • Select Private Echo from the voice assistants list.

You should see Home Assistant for Conversation agent. You should also see Speech-to-Phrase for Speech-to-text plus Piper for Text-to-speech.

  • Open the Debug view.
  • Choose Run text pipeline.
  • Enter turn off Workshop Light.

Before you select Run, do you think Workshop Light will stay on or switch off?

  • Select Run.

The Debug result confirms that Private Echo processed the text command. The same pipeline now has local speech providers ready for future audio.

  • Return to the Home Assistant dashboard from earlier.

You should see the Workshop Light toggle switched off. You have closed the missing-engine gap.

Workshop Light Still On?

  • Confirm that the Debug input says turn off Workshop Light exactly.
  • Confirm that Home Assistant remains selected as the Conversation agent.
  • Confirm that the Workshop Light helper remains exposed to Assist.

Help me troubleshoot a Private Echo Debug command that does not turn off Workshop Light.

Your local Assist pipeline can now process a command that changes Workshop Light. Next, you will create a safety brief before any Echo hardware or firmware work begins.

Create the Echo Safety Brief

Your local Private Echo pipeline gives the future Echo a working destination inside Home Assistant.

The later conversion can unlock the bootloader. It can also write the system partition.

In this step, Claude Chat helps translate the canonical requirements into a reviewed safety brief. No Echo modification occurs.

In this step, get ready to:
  • Ask Claude Chat to explain the canonical source under strict read-only limits.
  • Correct the draft against the canonical device page.
  • Record safe follow-on paths for Intel Macs and Apple silicon Macs.
Gather the canonical safety evidence

The canonical echo-dot-assist device page defines the requirements for this Echo model. Claude Chat will explain only the excerpts you provide.

Keep This Step Read-Only

The Echo stays untouched throughout this step.

  • Run no rooting command during this project.
  • Make no hardware change to the Echo.
  • Open the canonical D9N29T device page in your Mac browser.
  • Locate the Requirements heading on the device page.
  • Select only the text under the Requirements heading.
  • Copy the selected text.
  • Open claude.ai in another browser tab.

The Requirements excerpt is now on your clipboard.

  • Paste the Requirements excerpt into the Claude Chat message field.
  • Return to the canonical device page.
  • Locate the USB access heading.
  • Select only the text under the USB access heading.
  • Copy the selected text.
  • Return to Claude Chat.
  • Paste the USB access excerpt below the Requirements excerpt.
  • Add the safety request beneath both excerpts by pasting this prompt:
Explain each requirement in plain language. Separate read-only checks from actions that can write to the Echo. List every point where I must stop instead of guessing. Do not tell me to run any command.

Why This Prompt Is Safe

  • The first sentence asks for a plain-language explanation.
  • The second sentence creates a boundary between inspection and modification.
  • The third sentence turns uncertainty into a stop condition.
  • The final sentence prevents the draft from becoming a procedure.
  • Send the message in Claude Chat.
  • Read the response for a clear separation between read-only checks and writing actions.

You should see a plain-language draft that identifies where you must stop instead of guessing.

Did Claude Include Commands?

  • Run none of the commands included in the response.
  • Resend the same source excerpts with the exact safety request above if the read-only boundary is missing.

Help me keep the Claude response read-only.

Save and correct the source-backed brief

Claude's response is a working draft. The canonical device page remains the source of truth.

  • Create a blank document in a document app on your Mac.
  • Copy Claude's response from the browser.
  • Paste the response into the blank document.
  • Save the document with the title Echo Conversion Safety Brief.
  • Set the Echo model in the brief to D9N29T.
  • Set the required firmware in the brief to Fire OS 6574.1.

You should now see both exact identifiers in the saved brief.

  • Write a host requirement that requires Linux running on the PC itself while excluding macOS and virtual machines.
  • Add a warning to keep the Echo away from Amazon updates during the later procedure.
  • Save the corrected document.

The brief now anchors its device requirements to the canonical page.

Add the stop conditions and future computer path

A safety gate must turn uncertainty into a clear reason to stop. It must also define a suitable computer path before the later conversion begins.

  • Add an explicit condition requiring you to stop instead of guessing whenever a requirement cannot be confirmed.
  • Add an explicit instruction that no rooting command is run during this project.

Those two statements make the current project boundary visible in the brief.

  • Add a statement that an Intel Mac may be evaluated with a Linux live environment in the next project.
  • Add a statement that an Apple silicon Mac should use a borrowed native-Linux computer.
  • Add a condition that avoids internal-storage changes on an Apple silicon Mac during this beginner project.

Why the Mac Paths Differ

An Intel Mac may be suitable for evaluating a Linux live environment in the follow-on project.

Apple silicon hardware does not support booting purely from USB storage without internal-storage changes. A borrowed native-Linux computer keeps those changes outside this beginner project.

  • Save the completed Echo Conversion Safety Brief.

Before you inspect the saved document, consider whether every risky assumption now has a stop condition.

  • Confirm the brief identifies model D9N29T.
  • Confirm the brief requires Fire OS 6574.1.
  • Confirm the brief requires Linux to run directly on the PC while excluding macOS and virtual machines.
  • Confirm the brief warns against Amazon updates during the later procedure.
  • Confirm the brief requires you to stop instead of guessing.
  • Confirm the brief states that no rooting command is run in this project.
  • Confirm the brief records the Intel Mac evaluation path.
  • Confirm the brief records the borrowed native-Linux path for Apple silicon.

Strong work. Your saved brief now blocks guesswork before the risky Echo phase.

Secret mission

Teach Assist a Private Workshop Command

Give your private control center a phrase that means something specific in your workshop. The finished automation toggles Workshop Light and returns a custom confirmation through Assist.

Clean Up Your Resources

Clean Up Your Resources

Your private control center runs inside a local Home Assistant OS virtual machine in VirtualBox. It creates no paid cloud resource or ongoing financial cost.

Decide whether to keep your resources running, pause them to come back later, or delete them entirely.

Resources you used:

  • The Home Assistant VirtualBox VM with its Home Assistant OS installation.
  • The Speech-to-Phrase and Piper apps inside the VM.
  • The Wyoming integrations and the Private Echo Assist pipeline inside the VM.
  • The Workshop Light Toggle helper inside the VM.
  • The Private Workshop Command automation inside the VM.

Keep everything running

No action is needed. Choose this if you plan to continue experimenting with your private smart-home control center.

  • Leave the Home Assistant VM running in VirtualBox.
  • Keep the Mac powered on whenever you need Home Assistant to remain available.
  • Return to http://homeassistant.local/ when you are ready to continue.

Your dashboard, Workshop Light, local voice pipeline, and private-workshop automation remain available.

Pause - I'll come back to this later

Shut down the VM to free the memory and processor capacity it uses. Keep its files so your Home Assistant configuration is ready for next time.

  • Use Home Assistant's system controls to shut down the installation.
  • Wait for the Home Assistant shutdown to complete.
  • In VirtualBox, stop the Home Assistant VM.
  • Start the same Home Assistant VM in VirtualBox when you are ready to resume.

You'll see the VM stopped in VirtualBox while its configuration remains stored in the VM files.

Delete - I don't want to use this again

This is a permanent local reset. The Echo itself remains untouched because no modification occurred.

  • Use Home Assistant's system controls to shut down the installation.
  • Wait for the Home Assistant shutdown to complete.
  • In VirtualBox, select the Home Assistant VM.
  • Remove the selected VM from VirtualBox.
  • Choose the option that deletes the VM's files.

Deleting those files also removes Home Assistant OS, Speech-to-Phrase, Piper, the Wyoming integrations, the Private Echo pipeline, the Workshop Light helper, and the private-workshop automation.

You'll know cleanup is complete when Home Assistant no longer appears in VirtualBox.

Nice Work!

Nice Work!

You made it! Your Mac now runs a private Home Assistant control center in VirtualBox with local voice services. Your Echo remains untouched behind a written safety gate.

You've learned how to:

  • Run a local Home Assistant OS control center in a bridged VirtualBox VM. Control the Workshop Light through typed Assist commands.
  • Resolve the missing-provider obstacle with Speech-to-Phrase plus Piper in the Private Echo pipeline. Confirm the result through the Debug text pipeline.
  • Create a Claude-reviewed Echo Conversion Safety Brief with the D9N29T requirements. Protect the Echo with explicit stop conditions before any modification.
  • Secret Mission: Build a custom sentence automation for make the workshop private. Verify that Workshop Light changes state. Confirm that Assist returns Workshop mode changed.

Ready to quiz yourself?