Skip to content
Get started

ESPHome 2025.5.0: Voice assistant updates and packet transport

ESPHome 2025.5.0: Voice assistant updates and packet transport
Jesse Hills
Jesse Hills
ESPHome Developer
PM2​0​0​5​ ​Sensor
Mapping
Packet ​Transport
Syslog
Sound ​Level
MIPI ​SPI ​Displays

Voice Assistant

Section titled “Voice Assistant”

Home Assistant Voice Preview Edition is Home Assistant’s first official voice assistant hardware. It was designed from the ground up to give the best experience possible with Assist, Home Assistant’s built-in voice assistant software. With advanced mics and audio processing, it can run fully locally in the home. Voice Preview Edition was built to run ESPHome, and like any ESPHome device, can be fully customized.

We have implemented all of the new features developed for the Voice Preview Edition! The voice assistant component supports starting conversations from Home Assistant and continuing conversations without needing to reactivate the wake word. Additionally, microWakeWord supports enabling and disabling models via yaml actions. The voice assistant component, through Home Assistant, can also change the active on-device wake word. If you use our prebuilt voice assistant firmwares, then you will receive an update notification automatically in Home Assistant adding support for these new features!

If you have taken control or created your own voice assistant configuration, you may need to adjust them to support these new features. Please see the newly updated examples in our wake word voice assistant firmware repository to get started.

If you have taken control of your Home Assistant Voice PE and want to get all of the new changes and don’t want to continue managing your own YAML file, you can re-install the latest firmware from the web installer. We also provide OTA updates for the device that you can directly install from within Home Assistant.

It is now easier to configure and use microphones. Multiple components can simultaneously read from one microphone. Each individual component using a microphone can also set their own specific gain level or specify which microphone channel to use. Microphone settings are validated at compilation, and most settings are converted, if necessary, without any extra configuration. Your existing YAML files should still build without any adjustments. However, if your microphone was previously configured to use 32 bits per sample, then you should add a gain factor of 4 to match ESPHome’s previous behavior. For example:

voice_assistant:
  microphone:
    microphone: echo_microphone
    gain_factor: 4

Packet Transport

Section titled “Packet Transport”

The Packet Transport has been added to allow ESPHome devices to directly communicate with each other. There are two platforms available in this release: Uart and Udp.

HTTP Request

Section titled “HTTP Request”

The Http Request actions have had the headers config variable renamed to request_headers in order to differentiate it from the new collect_headers config variable allowing you to collect and access the response headers in automations.

ATM90E32

Section titled “ATM90E32”

For energy meters that use the atm90e32 energy metering chip, like from CircuitSetup, and Gelidus Research, support was added for semi-automatic gain & offset calibration, and status messages for each phase & frequency. An apparent power sensor was also added.

Release 2025.5.1 - May 27

Section titled “Release 2025.5.1 - May 27”

Release 2025.5.2 - June 3

Section titled “Release 2025.5.2 - June 3”

Full List of Changes

Section titled “Full List of Changes”

For the complete list of every merged pull request in this release, see the full 2025.5.0 changelog.

Tags:

Comments