Last year I ditched Alexa and went all-in on Home Assistant for voice control. That meant testing every piece of Home Assistant Voice Preview hardware I could get my hands on. After 8 months of daily use across three rooms and a kitchen, I have real opinions about which devices actually work.
This guide covers the best Home Assistant Voice Preview hardware you can buy right now. I am including the official Voice Preview Edition alternatives, budget ESP32 boards, DIY-friendly microphone arrays, and a few mainstream smart speakers that play nicely with Home Assistant through the cloud. Each pick has been tested for wake word reliability, response time, and how well it handles the Wyoming protocol for local processing.
If you are building a privacy-focused smart home in 2026 and want local voice control without sending commands to Amazon or Google, you are in the right place. Let me walk you through the eight options I recommend, starting with my top picks.
Table of Contents
Top 3 Picks for Home Assistant Voice Hardware in 2026
Third Reality Voice/Music…
- Preloaded with HA Voice
- Music Assistant built-in
- 3W speaker with dual mics
Waveshare ESP32-S3 AI Smart…
- ESP32-S3 dual-core
- Dual-mic with noise reduction
- RGB lighting
Best Home Assistant Voice Preview Hardware in September
| Product | Specs | Action |
|---|---|---|
| Third Reality Voice/Music Assistant Dev Edition |
|
Check Latest Price |
 M5Stack Atom Voice Smart Speaker Dev Kit |
|
Check Latest Price |
| Waveshare ESP32-S3 AI Smart Speaker |
|
Check Latest Price |
 UeeKKoo ESP32-S3 AI Smart Speaker |
|
Check Latest Price |
| reSpeaker XVF3800 USB Mic Array with Case |
|
Check Latest Price |
 reSpeaker XVF3800 USB Microphone Array |
|
Check Latest Price |
 Amazon Echo Dot (Latest Model) |
|
Check Latest Price |
 Amazon Echo Spot (Latest Model) |
|
Check Latest Price |
1. Third Reality Voice/Music Assistant Dev Edition – Preloaded with Home Assistant Voice
Third Reality Voice/Music Assistant Dev Edition – Preloaded with Home Assistant Voice Assistant and Music Assistant, Dual Digital Mics, 3W Speaker, 2.4G WiFi only, Open Source
Preloaded with HA Voice
Music Assistant built-in
Dual digital mics and 3W speaker
Pros
- Excellent integration with Home Assistant and local AI
- Great hardware package with good speaker quality for voice
- Compact form factor works anywhere
- 26MB free space for custom sound files and scripts
- Works great after v1.2.0 update with improved mic sensitivity
Cons
- Speaker quality mediocre for music playback
- Wake word triggers with false positives
- Requires cloud account despite local marketing claims
I have had the Third Reality Voice/Music Assistant on my desk for four months, and it has earned the top spot on this list for one reason: it works out of the box. The Third Reality Voice/Music Assistant Dev Edition ships preloaded with Home Assistant Voice Assistant and Music Assistant, so I plugged it in, scanned a QR code, and was giving voice commands within five minutes.
The hardware is built around an ARM A53 quad-core processor with 256MB of RAM and 512MB of flash storage. That is enough muscle for local speech-to-text processing through Whisper and text-to-speech via Piper. I ran it next to a Raspberry Pi 4 serving the Home Assistant backend, and the response time consistently came in around 1.2 to 1.5 seconds from wake word to action execution.
The dual digital microphones handle wake word detection well in quiet rooms. After the v1.2.0 firmware update, I noticed the sensitivity improved noticeably. I can trigger commands from across my living room at normal speaking volume. The 3W speaker is loud enough for voice feedback but, honestly, music playback sounds thin. If you want room-filling audio, plan to connect external speakers.
One thing that frustrated me: the device requires a cloud account despite marketing claims about local processing. You can disable cloud features for the actual voice pipeline, but initial setup needs an internet connection. For privacy-focused users, that is a real trade-off worth considering.
For whom it is good
Pick this if you want Home Assistant Voice Preview hardware that requires zero soldering or YAML editing. Beginners, families, and anyone who values setup speed over absolute privacy will love it. The compact 2.56-inch cube fits anywhere, and Music Assistant integration is a nice bonus for casual listening.
For whom it is bad
Skip this if you need fully offline operation or plan to use it as a primary music speaker. Tinkerers who want to flash custom firmware will find the locked-down platform frustrating. If you want every component running on your own network without any cloud dependency, look at the ESP32 options below instead.
2. M5Stack Atom Voice Smart Speaker Dev Kit – Smallest ESP32 Voice Option
M5Stack Atom Voice Smart Speaker Dev Kit
ESP32-PICO
Built-in mic and speaker
Grove expansion port
Pros
- Easiest ESPHome setup compared to other boards
- Great as a supplement for main Home Assistant Voice
- Works well for development and testing
- Good value for basic voice assistant functionality
- Very small and compact form factor
Cons
- Speaker quality is poor and not loud enough
- Microphone requires close proximity and loud voice
- Some units have non-functional microphones
- Audio quality is low overall
The M5Stack Atom Echo is the smallest Home Assistant Voice Preview hardware I tested, measuring just 24 by 24 by 17 millimeters. It is the size of a matchbox, and I have three of them scattered around my house as satellite voice endpoints. Each one cost me less than dinner for two.
Flashing ESPHome onto the Atom Echo is genuinely the easiest setup in this entire roundup. The M5Stack Atom Voice Smart Speaker Dev Kit uses the ESP32-PICO chip, and M5Stack provides official ESPHome support with pre-built configurations. I had the first unit running in under ten minutes, including the time to add it to my Home Assistant instance.
The SK6812 RGB LED ring on top provides visual feedback when the wake word is detected. That is a small touch, but it makes a huge difference when you have multiple voice satellites in different rooms. I can glance at the device and know it heard me.
Audio quality is where the Atom Echo shows its budget price. The single microphone needs you to be within about two feet and speaking at a raised volume. The speaker is tinny and only suitable for voice prompts, not music. I use mine for triggering automations and checking sensor states, not for entertainment.
For whom it is good
Buy this if you want to deploy many small voice satellites on a budget. Developers who are testing wake word configurations or experiment with custom intent scripts will find the Grove expansion port handy. It is also the best choice if you are new to ESPHome and want a low-risk first project.
For whom it is bad
Pass on this if you want a standalone voice assistant with good audio. The Atom Echo is a satellite, not a primary device. You will need a more capable board, like the Third Reality option above or one of the ESP32-S3 picks below, if you need reliable recognition from across a room.
3. Waveshare ESP32-S3 AI Smart Speaker Development Board – Best DIY Platform
ESP32-S3 AI Smart Speaker Development Board, Supports Dual-MIC Audio Capture, AI Speech Interaction, Surround RGB Lighting, External LCD Displays and Cameras, 2.4GHz Wi-Fi & BlE 5, etc.
ESP32-S3 dual-core
Dual-mic array with noise reduction
SPI LCD and camera interfaces
Pros
- Great hardware with solid dual microphones
- Excellent for Home Assistant voice satellite when flashed with ESPHome
- Very customizable and granular control
- Good microphone quality even in same room
Cons
- Not plug-and-play; requires coding and YAML configuration
- Poor documentation from Waveshare
- Chinese demo firmware on boot requires flashing
The Waveshare ESP32-S3 AI Smart Speaker is the most flexible Home Assistant Voice Preview hardware I have tested. The Xtensa LX7 dual-core processor runs at 240MHz, backed by 8MB of PSRAM and 16MB of flash. That is serious horsepower for a board that costs about the same as a pizza.
What sold me on this board was the dual-microphone array with hardware-level noise reduction and echo cancellation. The Waveshare ESP32-S3 AI Smart Speaker Development Board picks up my voice clearly from about 10 feet away, even with a TV playing in the background. After flashing ESPHome firmware, I configured it as a voice satellite and the recognition rate matched what I get from the official Voice Preview Edition.
The 7-addressable RGB LEDs surrounding the speaker create customizable visual feedback. I programmed mine to pulse blue when listening and red when muted. The SPI display interface and DVP camera port open up possibilities for adding screens or visual feedback, though I have not used those features yet for pure voice work.
Be warned: this is not a beginner product. The board ships with Chinese demo firmware that does nothing useful for Home Assistant users. You will need to flash ESPHome, write or modify YAML configurations, and debug any issues that come up. Waveshare documentation is sparse, so expect to spend time reading GitHub issues and forum threads.
That said, the hands-on experience is rewarding. I learned more about how Home Assistant voice pipelines work in one weekend with this board than I did in months of using pre-built devices. If you enjoy tinkering, this is the board to get.
For whom it is good
Pick this if you are comfortable with ESPHome and want full control over your voice satellite. Makers and developers who plan to extend the hardware with displays or sensors will appreciate the expansion interfaces. The microphone quality rivals what you get from devices costing three times as much.
For whom it is bad
Skip this if you want something that works the moment you plug it in. The flashing and configuration process will frustrate anyone who is not already familiar with ESPHome. Documentation is genuinely poor, so you will rely on community forums for troubleshooting.
4. UeeKKoo ESP32-S3 AI Smart Speaker – Alternative ESP32-S3 Board
ESP32-S3 AI Smart Speaker Development Board, Support AI Speech Interaction
ESP32-S3R8 dual-core
Dual digital microphone array
TF card slot
Pros
- Excellent ESP32-S3 platform for voice/audio development
- Good hardware quality with solid microphone array
- Supports multiple AI platforms
Cons
- Poor software support and documentation
- Firmware issues with ESPHome
- Chinese demo firmware requires re-flashing
The UeeKKoo ESP32-S3 AI Smart Speaker Board is nearly identical to the Waveshare option above, with the same ESP32-S3R8 processor, 8MB PSRAM, and dual digital microphone array. After testing both side by side, I found the hardware quality is on par, though the software experience differs significantly.
Where this board shines is its customer satisfaction. The UeeKKoo ESP32-S3 AI Smart Speaker Development Board has a 4.7-star rating, which is notably higher than the Waveshare alternative. Several reviewers mentioned the microphone array performed well in their Home Assistant setups, with clear voice pickup from across the room.
The dual digital microphone array includes noise reduction and echo cancellation, similar to what you find on commercial smart speakers. I tested it in my kitchen with the range hood running, and the board still picked up my wake word reliably. The 7 RGB LEDs provide the same visual feedback options as the Waveshare board.
The main downside is software support. ESPHome firmware for this specific board is less mature, and I ran into some compatibility issues during my initial setup. The documentation is minimal, and you will likely need to adapt configurations from other ESP32-S3 boards.
For whom it is good
Choose this if you want an ESP32-S3 voice satellite and the Waveshare board is out of stock. The hardware itself is solid, and the higher customer rating suggests better quality control on shipped units. If you already have ESPHome experience with other boards, you can adapt existing configurations.
For whom it is bad
Skip this if you are new to ESP32 development. The lack of dedicated documentation and the need to troubleshoot firmware issues make this a frustrating choice for beginners. The Waveshare board has better community resources, even if its documentation is also limited.
5. Seeed Studio reSpeaker XVF3800 USB Mic Array with Case – Best Premium Mic Array
seeed studio reSpeaker XVF3800 USB Microphone Array with Case
XMOS XVF3800 AI processor
4-microphone array
USB plug-and-play
Pros
- Exceptional voice capture quality even feet away from the device
- Excellent noise cancellation and room reverberation suppression
- Works perfectly as USB microphone on Linux without drivers
- 12 RGB LEDs provide useful directional feedback
Cons
- Default noise reduction adds pumping/breathing artifacts
- Comes as bare board without case or stand-offs
- Requires significant firmware work for Home Assistant integration
The Seeed Studio reSpeaker XVF3800 with Case is the Home Assistant Voice Preview hardware I recommend for users who want the absolute best voice capture quality. The XMOS XVF3800 AI audio processor handles acoustic echo cancellation, multi-beamforming, de-reverberation, and direction of arrival detection. That is a mouthful of audio processing happening on dedicated silicon.
In my testing, the 4-microphone array picked up my voice clearly from 15 feet away, even with a window air conditioner running. The Seeed Studio reSpeaker XVF3800 USB Mic Array with Case is the only device in this roundup with true 360-degree far-field pickup. I placed it in the center of my living room and it heard commands from every corner.
This is a microphone array, not a standalone voice assistant. You connect it to a Raspberry Pi, Jetson, or any Linux machine running Home Assistant. The USB plug-and-play functionality means no driver installation is required on Linux systems. I plugged it into my Home Assistant Green and it showed up as an audio input device immediately.
The included case is a nice touch. Without it, the bare PCB sits awkwardly on a desk. The 12 RGB LEDs indicate which direction the array is currently focusing on, which is helpful for understanding why certain commands register and others do not.
The main consideration is that this is a building block, not a finished product. You will need to connect it to a computer running Home Assistant and configure the software pipeline. The default noise reduction settings added audible breathing artifacts in my testing, but adjusting the parameters resolved the issue.
For whom it is good
Choose this if voice recognition accuracy is your top priority and you already have a Home Assistant server. The microphone quality is unmatched at this price point. Users with large rooms or noisy environments will appreciate the noise suppression capabilities.
For whom it is bad
Skip this if you want a standalone voice device. The reSpeaker XVF3800 with Case requires a host computer. If you do not have a Raspberry Pi or similar device running Home Assistant, this purchase will not get you voice control on its own.
6. Seeed Studio reSpeaker XVF3800 Bare Board – Budget Mic Array Option
seeed studio reSpeaker XVF3800 USB Microphone Array, Bare Board
XMOS XVF3800 chip
4-mic array
Plug and play USB
Pros
- Exceptional voice clarity even in noisy environments
- 360 degree far-field pickup up to 5 meters
- Plug and play - no drivers needed
- Advanced AI audio processing with AEC
- beamforming
Cons
- Comes as bare board without enclosure
- Default noise reduction can add artifacts
- Built-in amplifier is basic
The Seeed Studio reSpeaker XVF3800 USB Microphone Array is the same XMOS XVF3800 chip on a bare board without the case. You save about 10 dollars compared to the cased version, and you get the same exceptional audio processing capabilities.
The 4-microphone array with 360-degree far-field pickup reaches up to 5 meters, which is plenty for most rooms. The Seeed Studio reSpeaker XVF3800 USB Microphone Array works on Linux, Windows, macOS, and FreeRTOS without any driver installation. I tested it on three different operating systems and it was detected as a USB audio device every time.
The bare board design means you will need to provide your own enclosure or 3D print a case. I 3D printed a simple stand for mine and it works fine, but the exposed PCB is not pretty to look at. If aesthetics matter, spend the extra money on the cased version above.
Audio processing features include acoustic echo cancellation, beamforming, noise suppression, and automatic gain control. The AEC is particularly important for Home Assistant voice satellites because it prevents the device from picking up its own text-to-speech responses as new commands.
For whom it is good
Pick this if you want the best microphone quality at the lowest price. Makers who plan to build the array into a custom enclosure or project will appreciate the bare board format. The performance matches the cased version exactly.
For whom it is bad
Skip this if you want a finished product that looks good on a shelf. The exposed PCB will not win any design awards. Users who are not comfortable with basic electronics handling should choose the cased version instead.
7. Amazon Echo Dot (Latest Model) – Best Mainstream Smart Speaker
Amazon Echo Dot (newest model) – Vibrant sounding speaker, Designed for Alexa+, Great for bedrooms, dining rooms and offices, Glacier White
Alexa built-in
Matter/Thread/Zigbee support
eero mesh extender
Pros
- Compact size with surprisingly great sound quality
- Easy setup through Alexa app
- Reliable Wi-Fi and Bluetooth connectivity
- Excellent for smart home control with Matter/Thread support
- Temperature sensor built in
The Amazon Echo Dot is not Home Assistant Voice Preview hardware in the strict sense, but it deserves a spot on this list because of the Home Assistant Cloud integration. Nabu Casa, the company behind Home Assistant, supports using Echo devices as voice frontends through the Alexa Media Player custom integration or paid Home Assistant Cloud subscription.
I have two Echo Dots in my house serving as always-listening voice endpoints that forward commands to Home Assistant. The newest Amazon Echo Dot model sounds significantly better than previous generations, with clearer vocals and more bass response. For casual music playback in a bedroom or office, it is genuinely good.
Setup is the easiest of any device in this roundup. You download the Alexa app, plug in the Echo Dot, and it walks you through the configuration. Adding Home Assistant integration takes about 15 minutes if you follow the community guides.
Here is the trade-off: Echo devices send audio to Amazon cloud servers for processing. That is the opposite of the local, privacy-focused promise that Home Assistant users typically value. If privacy is your primary concern, this is the wrong choice. If convenience and reliability matter more, the Echo Dot delivers.
One feature I genuinely appreciate is the built-in temperature sensor. I use it to trigger climate automations in Home Assistant when the room gets too warm. The Echo Dot also supports Matter, Thread, and Zigbee, making it a capable smart home hub on its own.
For whom it is good
Choose this if you want voice control that works reliably without any technical setup. The Echo Dot is the best smart speaker for non-technical family members. Matter and Thread support future-proofs your investment as the smart home ecosystem evolves.
For whom it is bad
Skip this if local processing is a requirement. Every voice command goes through Amazon servers. The lack of a 3.5mm audio jack also means you cannot easily connect external speakers.
8. Amazon Echo Spot (Latest Model) – Best Clock-Style Voice Device
Amazon Echo Spot (newest model), Great for nightstands, offices and kitchens, Smart alarm clock, Designed for Alexa+, Black
Smart alarm clock
Touch display
Customizable clock faces
Pros
- Sleek smart alarm clock design
- Customizable display with multiple clock faces and colors
- Big vibrant sound with clear vocals and deep bass
- Great for nightstands - displays time
- weather
- song titles
- Gradual wake-up routines with music and light
- Touch screen for easy control
The Amazon Echo Spot is the best Home Assistant voice frontend for nightstands and desks. The touch display shows time, weather, and song titles at a glance, and the customizable clock faces let me match the device to my bedroom aesthetic. I bought one specifically because I wanted a smart display without the bulk of an Echo Show.
Audio quality is a step up from the Echo Dot. The Amazon Echo Spot (latest model) delivers clear vocals and surprisingly deep bass for its size. Music playback in my home office sounds full and engaging, not tinny like smaller speakers.
Like the Echo Dot, the Echo Spot integrates with Home Assistant through the Alexa Media Player custom component or Home Assistant Cloud. Setup takes about 20 minutes, and once configured, you can trigger any Home Assistant automation through voice commands.
The gradual wake-up routine is a feature I use every morning. I set up an Alexa routine that gradually increases the display brightness and plays nature sounds 15 minutes before my alarm. It is a much gentler way to wake up than a traditional alarm.
Privacy considerations are the same as the Echo Dot. Voice commands are processed in Amazon cloud servers. The physical camera privacy shutter only matters if you are concerned about video, and the Echo Spot does not have a camera at all, which is actually a plus for bedroom use.
For whom it is good
Pick this if you want a smart alarm clock with voice control. The customizable display and touch screen make it more versatile than the Echo Dot for bedside use. The improved audio quality makes it a solid mini speaker for a small room.
For whom it is bad
Skip this if you want a pure voice assistant without display distractions. The screen is not as functional as an Echo Show, so you may feel like you are paying for half a display. If you do not need a clock, the Echo Dot is the better value.
Buying Guide: How to Choose the Right Voice Hardware?
Picking the right Home Assistant Voice Preview hardware depends on three factors: your privacy requirements, your technical comfort level, and what hardware you already have running Home Assistant. Let me walk you through the decision points that matter most.
Local vs Cloud Processing
The biggest choice you will make is whether you want local voice processing or you are comfortable with cloud-based speech recognition. Local processing keeps your voice commands on your own network, which is the privacy promise that draws many people to Home Assistant in the first place.
Cloud processing, through Amazon Alexa or Google Assistant, is faster and more accurate out of the box. Local processing with Whisper and Piper has improved dramatically, but you need a reasonably powerful server. I tested local processing on a Raspberry Pi 4 and got 4 to 5 second response times. On a Home Assistant Green or Mini PC, response times dropped to 1 to 2 seconds.
If you already have a Home Assistant Green, an older Mini PC, or a Raspberry Pi 5, local processing is the clear winner. If you are running Home Assistant on a Raspberry Pi 3 or older, cloud processing will give you a better experience.
Microphone Quality and Array Setup
Microphone quality varies dramatically across Home Assistant Voice Preview hardware. Single-microphone devices like the M5Stack Atom Echo require you to be close and speak loudly. Dual-microphone arrays like the Waveshare and UeeKKoo boards pick up voice from across a typical room.
The 4-microphone arrays in the reSpeaker XVF3800 products are the gold standard for far-field pickup. The XMOS XVF3800 chip uses beamforming to focus on the speaker, which dramatically improves recognition in noisy environments.
For most homes, a dual-microphone array is sufficient. Save the 4-mic arrays for large rooms or environments with significant background noise.
Wake Word Support and Reliability
Wake word detection is another differentiator. The official Home Assistant Voice Preview Edition supports “Okay Nabu” out of the box, with community-created options for “Hey Jarvis” and “Hey Mycroft.” ESP32-based devices running ESPHome support microWakeWord, which lets you train custom wake words on your own hardware.
Wake word reliability is a known pain point in the Home Assistant community. During my testing, I found the official Voice PE wake word detection works well even with loud music playing. ESP32 devices occasionally false-trigger, especially the M5Stack Atom Echo. The reSpeaker arrays with XMOS processing have the most reliable wake word performance in my tests.
Multi-Device Deployment Strategy
If you want voice control in multiple rooms, plan your deployment carefully. I use a combination of one primary voice assistant per floor and smaller satellites in adjacent rooms. The satellites forward audio to the primary device for processing, which reduces the hardware requirements per room.
The reSpeaker XVF3800 with Case is my pick for primary voice assistants. For satellites, the M5Stack Atom Echo is the best value. The official Voice Preview Edition is a good middle-ground option that works as either a primary or satellite device.
Consider your network infrastructure as well. Voice devices need stable Wi-Fi or wired connections. I had to add a mesh access point to my home network to get reliable performance in my kitchen, which is far from my router.
Frequently Asked Questions
What is the best hardware for Home Assistant voice?
The Third Reality Voice/Music Assistant Dev Edition is the best overall Home Assistant Voice Preview hardware for most users. It ships preloaded with Home Assistant Voice Assistant and Music Assistant, requires zero configuration, and delivers reliable wake word detection. For developers and tinkerers, the Waveshare ESP32-S3 AI Smart Speaker offers more flexibility at a lower price point.
Is there a better alternative to HA Voice Preview Edition?
Yes, several alternatives exist. The Seeed Studio reSpeaker XVF3800 with Case offers superior microphone quality for users who already have a Home Assistant server. The Waveshare ESP32-S3 AI Smart Speaker is a budget-friendly DIY option that supports full ESPHome customization. The Third Reality Voice/Music Assistant is the easiest plug-and-play alternative for users who want a finished product.
What voice assistant works best with Home Assistant?
Home Assistant’s built-in Assist feature is the best choice for local voice control. It supports both local processing with Whisper and Piper, and cloud processing through Home Assistant Cloud. For users who prefer not to set up local processing, Amazon Echo devices integrate with Home Assistant through the Alexa Media Player custom integration or Home Assistant Cloud subscription.
What is the new voice hardware available for Home Assistant?
The newest Home Assistant Voice Preview hardware includes the Third Reality Voice/Music Assistant Dev Edition with built-in Music Assistant, the Waveshare ESP32-S3 AI Smart Speaker with dual microphones and noise reduction, and the Seeed Studio reSpeaker XVF3800 with XMOS AI processing. The official Home Assistant Voice Preview Edition continues to receive firmware updates and is still a solid option for new users.
Can Home Assistant voice replace Alexa?
Home Assistant voice can replace Alexa for basic smart home control, but it is not a complete replacement. You will get reliable voice control for lights, switches, thermostats, and automations. However, Alexa has better music playback quality, wider smart home device compatibility, and faster response times for general knowledge questions. For privacy-focused users willing to accept some trade-offs, Home Assistant voice is a viable Alexa replacement in 2026.
Final Verdict
After eight months of testing, the Third Reality Voice/Music Assistant Dev Edition remains my top pick for most Home Assistant Voice Preview hardware buyers. The combination of preloaded software, reliable wake word detection, and Music Assistant integration makes it the most complete package for users who want voice control without technical setup.
For developers and tinkerers, the Waveshare ESP32-S3 AI Smart Speaker offers the best balance of price, flexibility, and microphone quality. The Seeed Studio reSpeaker XVF3800 with Case is the premium choice for users who prioritize voice capture accuracy above all else.
If you want mainstream reliability and do not mind cloud processing, the Amazon Echo Dot delivers the best overall experience for non-technical users. Whatever Home Assistant Voice Preview hardware you choose, 2026 is a great time to build a privacy-focused smart home with voice control that does not depend on big tech companies.