Home Assistant Voice PE
El volumen ha cambiado
AI Voice Assistant
Ask anything. Control everything.
Turn an ESPHome voice device into a natural voice interface for your Homey. Speak normally to control lights, thermostats, locks and everything else in your home, set timers and alarms, play music, and get quick answers to everyday questions. There is no rigid command syntax to learn, and the conversation stays open, so you can ask a follow-up without repeating the wake word. The assistant speaks your language — English, Dutch, German, French, Italian, Swedish, Norwegian, Spanish, Danish, Russian, Polish and Korean.
You choose the AI behind it: a low-latency cloud engine from OpenAI, Google or Mistral, or your own self-hosted services, so that no audio ever leaves your network. Extras such as web search, music and shopping lists can be switched on when you want them and stay out of the way when you don't.
Smart locks: locking by voice always works, but unlocking is off by default — anyone within earshot can talk to the device, so you must first enable "Allow unlocking by voice" in the app settings. Even then only one lock can be unlocked per command.
Supported devices, setup instructions, Flow cards and all the technical details are in the GitHub repository and the Homey Community topic linked below.
Dispositivos compatibles
Tarjetas Flow
Haz clic en el dispositivo de arriba para mostrar solo las tarjetas Flow relacionadas.
Cuando...
Home Assistant Voice PE
Encendido
Home Assistant Voice PE
Apagado
Home Assistant Voice PE
Heard something
Fires when speech-to-text finishes transcribing what was said to this device. Useful for debugging: pipe the text to the timeline or a logger to see exactly what the assistant heard.
Home Assistant Voice PE
Thinking
Fires while the assistant works on an answer: once for every tool it uses ("Using tool get_devices") and once with the final reply. Useful for debugging: pipe the text to the timeline or a logger to follow what the assistant is doing.
Home Assistant Voice PE
Reply audio is ready
Fires when the assistant's spoken reply has been rendered to an audio file, with a link a speaker can play, and for the app's own sound effects (the wake cue, an error, a missing API key, the greeting after pairing) — the "Is a sound effect" tag tells the two apart. Only fires when the device's "Reply audio" setting is set to send the reply to a Flow — use it to play the assistant's own voice on a Sonos or any other speaker that can play a URL. A reply link is only valid for a couple of minutes, so play it right away rather than storing it.
Home Assistant Voice PE
Timer cancelled
Fires when a running or ringing timer is cancelled (by voice, a flow, or being replaced).
Home Assistant Voice PE
Timer finished
Fires when a timer or alarm reaches zero and the device starts ringing.
Home Assistant Voice PE
Timer started
Fires when a timer or alarm starts counting down on the device.
M5Stack AtomS3R + Echo Base (ESPHome)
El volumen ha cambiado
M5Stack AtomS3R + Echo Base (ESPHome)
Encendido
M5Stack AtomS3R + Echo Base (ESPHome)
Apagado
M5Stack AtomS3R + Echo Base (ESPHome)
Heard something
Fires when speech-to-text finishes transcribing what was said to this device. Useful for debugging: pipe the text to the timeline or a logger to see exactly what the assistant heard.
M5Stack AtomS3R + Echo Base (ESPHome)
Thinking
Fires while the assistant works on an answer: once for every tool it uses ("Using tool get_devices") and once with the final reply. Useful for debugging: pipe the text to the timeline or a logger to follow what the assistant is doing.
M5Stack AtomS3R + Echo Base (ESPHome)
Reply audio is ready
Fires when the assistant's spoken reply has been rendered to an audio file, with a link a speaker can play, and for the app's own sound effects (the wake cue, an error, a missing API key, the greeting after pairing) — the "Is a sound effect" tag tells the two apart. Only fires when the device's "Reply audio" setting is set to send the reply to a Flow — use it to play the assistant's own voice on a Sonos or any other speaker that can play a URL. A reply link is only valid for a couple of minutes, so play it right away rather than storing it.
M5Stack AtomS3R + Echo Base (ESPHome)
Timer cancelled
Fires when a running or ringing timer is cancelled (by voice, a flow, or being replaced).
M5Stack AtomS3R + Echo Base (ESPHome)
Timer finished
Fires when a timer or alarm reaches zero and the device starts ringing.
M5Stack AtomS3R + Echo Base (ESPHome)
Timer started
Fires when a timer or alarm starts counting down on the device.
ReSpeaker XVF3800 (ESPHome)
El volumen ha cambiado
ReSpeaker XVF3800 (ESPHome)
Encendido
ReSpeaker XVF3800 (ESPHome)
Apagado
ReSpeaker XVF3800 (ESPHome)
Heard something
Fires when speech-to-text finishes transcribing what was said to this device. Useful for debugging: pipe the text to the timeline or a logger to see exactly what the assistant heard.
ReSpeaker XVF3800 (ESPHome)
Thinking
Fires while the assistant works on an answer: once for every tool it uses ("Using tool get_devices") and once with the final reply. Useful for debugging: pipe the text to the timeline or a logger to follow what the assistant is doing.
ReSpeaker XVF3800 (ESPHome)
Reply audio is ready
Fires when the assistant's spoken reply has been rendered to an audio file, with a link a speaker can play, and for the app's own sound effects (the wake cue, an error, a missing API key, the greeting after pairing) — the "Is a sound effect" tag tells the two apart. Only fires when the device's "Reply audio" setting is set to send the reply to a Flow — use it to play the assistant's own voice on a Sonos or any other speaker that can play a URL. A reply link is only valid for a couple of minutes, so play it right away rather than storing it.
ReSpeaker XVF3800 (ESPHome)
Timer cancelled
Fires when a running or ringing timer is cancelled (by voice, a flow, or being replaced).
ReSpeaker XVF3800 (ESPHome)
Timer finished
Fires when a timer or alarm reaches zero and the device starts ringing.
ReSpeaker XVF3800 (ESPHome)
Timer started
Fires when a timer or alarm starts counting down on the device.
ThirdReality Voice & Music Assistant
El volumen ha cambiado
ThirdReality Voice & Music Assistant
Encendido
ThirdReality Voice & Music Assistant
Apagado
ThirdReality Voice & Music Assistant
Heard something
Fires when speech-to-text finishes transcribing what was said to this device. Useful for debugging: pipe the text to the timeline or a logger to see exactly what the assistant heard.
ThirdReality Voice & Music Assistant
Thinking
Fires while the assistant works on an answer: once for every tool it uses ("Using tool get_devices") and once with the final reply. Useful for debugging: pipe the text to the timeline or a logger to follow what the assistant is doing.
ThirdReality Voice & Music Assistant
Button pressed
Fires when the physical button on top of the device is pressed.
ThirdReality Voice & Music Assistant
Reply audio is ready
Fires when the assistant's spoken reply has been rendered to an audio file, with a link a speaker can play, and for the app's own sound effects (the wake cue, an error, a missing API key, the greeting after pairing) — the "Is a sound effect" tag tells the two apart. Only fires when the device's "Reply audio" setting is set to send the reply to a Flow — use it to play the assistant's own voice on a Sonos or any other speaker that can play a URL. A reply link is only valid for a couple of minutes, so play it right away rather than storing it.
Xiaozhi AI
El volumen ha cambiado
Xiaozhi AI
Encendido
Xiaozhi AI
Apagado
Xiaozhi AI
Heard something
Fires when speech-to-text finishes transcribing what was said to this device. Useful for debugging: pipe the text to the timeline or a logger to see exactly what the assistant heard.
Xiaozhi AI
Thinking
Fires while the assistant works on an answer: once for every tool it uses ("Using tool get_devices") and once with the final reply. Useful for debugging: pipe the text to the timeline or a logger to follow what the assistant is doing.
Xiaozhi AI
Reply audio is ready
Fires when the assistant's spoken reply has been rendered to an audio file, with a link a speaker can play, and for the app's own sound effects (the wake cue, an error, a missing API key, the greeting after pairing) — the "Is a sound effect" tag tells the two apart. Only fires when the device's "Reply audio" setting is set to send the reply to a Flow — use it to play the assistant's own voice on a Sonos or any other speaker that can play a URL. A reply link is only valid for a couple of minutes, so play it right away rather than storing it.
Xiaozhi AI
Timer cancelled
Fires when a running or ringing timer is cancelled (by voice, a flow, or being replaced).
Xiaozhi AI
Timer finished
Fires when a timer or alarm reaches zero and the device starts ringing.
Xiaozhi AI
Timer started
Fires when a timer or alarm starts counting down on the device.
Y...
Home Assistant Voice PE
Está encendido
Home Assistant Voice PE
Is muted
Home Assistant Voice PE
A timer is running
True while a timer or alarm is counting down. A finished (ringing) timer does not count as running.
M5Stack AtomS3R + Echo Base (ESPHome)
Está encendido
M5Stack AtomS3R + Echo Base (ESPHome)
Is muted
M5Stack AtomS3R + Echo Base (ESPHome)
A timer is running
True while a timer or alarm is counting down. A finished (ringing) timer does not count as running.
ReSpeaker XVF3800 (ESPHome)
Está encendido
ReSpeaker XVF3800 (ESPHome)
Is muted
ReSpeaker XVF3800 (ESPHome)
A timer is running
True while a timer or alarm is counting down. A finished (ringing) timer does not count as running.
ThirdReality Voice & Music Assistant
Está encendido
ThirdReality Voice & Music Assistant
Is muted
Xiaozhi AI
Está encendido
Xiaozhi AI
A timer is running
True while a timer or alarm is counting down. A finished (ringing) timer does not count as running.
Entonces...
Home Assistant Voice PE
Configurar el volumen a %
Home Assistant Voice PE
Configurar volumen relativo %
Cambia el volumen con respecto al volumen actual del dispositivo.
Home Assistant Voice PE
Encender
Home Assistant Voice PE
Apagar
Home Assistant Voice PE
Encender o apagar
Home Assistant Voice PE
Silenciar el volumen
Home Assistant Voice PE
Activar el volumen
Home Assistant Voice PE
Activar el volumen o silenciar
Home Assistant Voice PE
Ask assistant Question. Output as text
Send an inquiry to the AI assistant, the response will be in a text variable
Advanced
Home Assistant Voice PE
Ask assistant Question. Output to speaker
Send an inquiry to the AI assistant, the response will play on this device
Home Assistant Voice PE
Cancel the timer
Cancel the running timer, or stop the timer/alarm currently ringing on the device.
Home Assistant Voice PE
Playback audio from Url
Play any (local or remote) web URL as long as it's flac format
Home Assistant Voice PE
Say text
Convert text to speech and play it
Home Assistant Voice PE
Start a Duration (seconds) second timer named Name
Start a countdown timer on the device. Replaces any timer already running (only one timer can exist at a time).
M5Stack AtomS3R + Echo Base (ESPHome)
Configurar el volumen a %
M5Stack AtomS3R + Echo Base (ESPHome)
Configurar volumen relativo %
Cambia el volumen con respecto al volumen actual del dispositivo.
M5Stack AtomS3R + Echo Base (ESPHome)
Encender
M5Stack AtomS3R + Echo Base (ESPHome)
Apagar
M5Stack AtomS3R + Echo Base (ESPHome)
Encender o apagar
M5Stack AtomS3R + Echo Base (ESPHome)
Silenciar el volumen
M5Stack AtomS3R + Echo Base (ESPHome)
Activar el volumen
M5Stack AtomS3R + Echo Base (ESPHome)
Activar el volumen o silenciar
M5Stack AtomS3R + Echo Base (ESPHome)
Ask assistant Question. Output as text
Send an inquiry to the AI assistant, the response will be in a text variable
Advanced
M5Stack AtomS3R + Echo Base (ESPHome)
Ask assistant Question. Output to speaker
Send an inquiry to the AI assistant, the response will play on this device
M5Stack AtomS3R + Echo Base (ESPHome)
Cancel the timer
Cancel the running timer, or stop the timer/alarm currently ringing on the device.
M5Stack AtomS3R + Echo Base (ESPHome)
Playback audio from Url
Play any (local or remote) web URL as long as it's flac format
M5Stack AtomS3R + Echo Base (ESPHome)
Say text
Convert text to speech and play it
M5Stack AtomS3R + Echo Base (ESPHome)
Start a Duration (seconds) second timer named Name
Start a countdown timer on the device. Replaces any timer already running (only one timer can exist at a time).
ReSpeaker XVF3800 (ESPHome)
Configurar el volumen a %
ReSpeaker XVF3800 (ESPHome)
Configurar volumen relativo %
Cambia el volumen con respecto al volumen actual del dispositivo.
ReSpeaker XVF3800 (ESPHome)
Encender
ReSpeaker XVF3800 (ESPHome)
Apagar
ReSpeaker XVF3800 (ESPHome)
Encender o apagar
ReSpeaker XVF3800 (ESPHome)
Silenciar el volumen
ReSpeaker XVF3800 (ESPHome)
Activar el volumen
ReSpeaker XVF3800 (ESPHome)
Activar el volumen o silenciar
ReSpeaker XVF3800 (ESPHome)
Ask assistant Question. Output as text
Send an inquiry to the AI assistant, the response will be in a text variable
Advanced
ReSpeaker XVF3800 (ESPHome)
Ask assistant Question. Output to speaker
Send an inquiry to the AI assistant, the response will play on this device
ReSpeaker XVF3800 (ESPHome)
Cancel the timer
Cancel the running timer, or stop the timer/alarm currently ringing on the device.
ReSpeaker XVF3800 (ESPHome)
Playback audio from Url
Play any (local or remote) web URL as long as it's flac format
ReSpeaker XVF3800 (ESPHome)
Say text
Convert text to speech and play it
ReSpeaker XVF3800 (ESPHome)
Start a Duration (seconds) second timer named Name
Start a countdown timer on the device. Replaces any timer already running (only one timer can exist at a time).
ThirdReality Voice & Music Assistant
Configurar el volumen a %
ThirdReality Voice & Music Assistant
Configurar volumen relativo %
Cambia el volumen con respecto al volumen actual del dispositivo.
ThirdReality Voice & Music Assistant
Encender
ThirdReality Voice & Music Assistant
Apagar
ThirdReality Voice & Music Assistant
Encender o apagar
ThirdReality Voice & Music Assistant
Silenciar el volumen
ThirdReality Voice & Music Assistant
Activar el volumen
ThirdReality Voice & Music Assistant
Activar el volumen o silenciar
ThirdReality Voice & Music Assistant
Ask assistant Question. Output as text
Send an inquiry to the AI assistant, the response will be in a text variable
Advanced
ThirdReality Voice & Music Assistant
Ask assistant Question. Output to speaker
Send an inquiry to the AI assistant, the response will play on this device
ThirdReality Voice & Music Assistant
Playback audio from Url
Play any (local or remote) web URL as long as it's flac format
ThirdReality Voice & Music Assistant
Say text
Convert text to speech and play it
Xiaozhi AI
Configurar el volumen a %
Xiaozhi AI
Configurar volumen relativo %
Cambia el volumen con respecto al volumen actual del dispositivo.
Xiaozhi AI
Encender
Xiaozhi AI
Apagar
Xiaozhi AI
Encender o apagar
Xiaozhi AI
Silenciar el volumen
Xiaozhi AI
Activar el volumen
Xiaozhi AI
Activar el volumen o silenciar
Xiaozhi AI
Ask assistant Question. Output as text
Send an inquiry to the AI assistant, the response will be in a text variable
Advanced
Xiaozhi AI
Ask assistant Question. Output to speaker
Send an inquiry to the AI assistant, the response will play on this device
Xiaozhi AI
Cancel the timer
Cancel the running timer, or stop the timer/alarm currently ringing on the device.
Xiaozhi AI
Playback audio from Url
Play any (local or remote) web URL as long as it's flac format
Xiaozhi AI
Say text
Convert text to speech and play it
Xiaozhi AI
Start a Duration (seconds) second timer named Name
Start a countdown timer on the device. Replaces any timer already running (only one timer can exist at a time).
Asistencia
¿Tienes algún problema con esta aplicación? Contacta con el desarrollador aquí.
Novedades
Versión 1.5.0 — New: the Custom pipeline can now run on Anthropic's Claude, with the model picked from a live list, and the cloud services for its speech and text stages — OpenAI, Groq, OpenRouter, DeepSeek — are now a dropdown that fills in the address and a working model for you. The pipeline settings are reorganised into one card per stage, and speech-to-text can be given hints (a context sentence and expected words) so short commands and names come back right. Important fix: adding a device could reboot your voice satellite on some ESPHome firmware versions, which showed up as pairing timing out or a device stuck Unavailable. An unavailable device now says which side is down — the device or the voice engine — and the Debug page shows the two separately, alongside a new Verbose logging switch for when you need to send in a log. Setting up Wi-Fi over Bluetooth now notices straight away if the device drops the link. Fixes a crash when a device was deleted, and the Home Assistant Voice PE now trims the wake word from the start of a command by default.
AI Voice Assistant
Esta aplicación puede administrar todo tu Homey.
AI Voice Assistant no es compatible con el Homey seleccionado.
AI Voice Assistant se instalará en Homey en breve.
Registro de cambios
¡Enviaremos un agradecimiento a Arve Bjørnerud!
Le haremos saber a Arve Bjørnerud que te encanta la aplicación. Quizás consideres apoyar a Arve Bjørnerud haciendo una donación.
Donar a Arve Bjørnerud
El desarrollador de AI Voice Assistant acepta donaciones.
¡Haz una donación para dar las gracias!
Se han enviado tus sugerencias
Avisaremos a Arve Bjørnerud para que esté al tanto de las sugerencias.
Africa
Asia
Europe
Belgium
Deutsch English
Bulgaria
Croatia
Cyprus
Czech Republic
Estonia
Finland
Greece
Hungary
Iceland
Ireland
Latvia
Lithuania
Malta
Moldova
The Netherlands
Portugal
Romania
Slovakia
Slovenia
Türkiye
Ukraine
United Kingdom
