#Debugging Continuous speech

1 messages · Page 1 of 1 (latest)

icy python
#

Recently got a Voice PE (and chucked out Google stuff). It's a fairly positive experience so far but I'm keen to get continuous speech working so the interaction feels more natural.
I've seen the latest release should support this so I wanted validate that I could expect it to be working and debug why not, if it should.

My setup is the following containers running on a bridge network happily talking to each other

I've attached the the ESPHome log from the PE when I say "knock knock" after the wake word.

Potentially I'm missing something obvious about how to enable this, or perhaps it's not supported in HA container. Happy to share any other details about config that are relevant.

image.png
frank jungle
icy python
#

ESPHome version 2025.3.2 compiled on Mar 28 2025, 14:31:19

#

that what we need?

frank jungle
#

looks like your not on latest

icy python
#

ohh. sounds promising

frank jungle
#

ensure that the firmware entity in configuration settings is enabled

#

if its enabled it should allow for it to get updates via the UI

icy python
#

hmm, not seeing it in the HA UI

frank jungle
#

in this section of the device page

icy python
frank jungle
#

ok i suggest a manual install then

#

plug the pe into your pc

#

this will allow you to install the latest firmware. you will have to enter your wifi creds again during this process

icy python
#

OK, is it worth trying to do a clean rebuild of the FW with esphome given that my config dosn't seem to be stuck to a specific version?

image.png
frank jungle
#

oh you took control?

#

are you planning on messing with the firmware?
if not then just do a clean install and leave it stock. theres no need to take control

icy python
#

just adopted it so i can/manage it (and all my other ESP32s) around the house - not intending to futz with the FW but its handy to keep SSIDs and stuff in one config file

frank jungle
#

i adopt alot of my devices but i dont with the VPE

#

your esphome is out of date

#

above you said your on
esphome/esphome:2025.3.2
latest is 2025.3.3

#

you could try updating that then rebuilding

icy python
#

alright i'll give it a go 👍

#

thanks for you help so far BTW

#

So I bumped the ESPHome version and rebuilt the firmware but no obvious difference so I'll have a go at the local install route.

frank jungle
#

you might have to delete and readd to HA after its flashed

#

again, i recomend not taking control. it makes life easier for updates for the VPE

icy python
#

Thanks! That's got it added to HA

image.png
#

So the question now is, do i have continuous speech.

frank jungle
#

indeed

icy python
#

hmm, seems like that's a no

#

should i be running "beta firmware" ?

frank jungle
#

no, the stock firmware should have it. it got added same day is the 2025.4 release

icy python
#

and I don't need to manually enable it? or set it in config anywhere 🤔

frank jungle
#

something of note: in your initial post the whos there was actually
whos there? EMOTE

#

i think the responce has to end in a ?

#

may have to rework your prompt to not use emotes

icy python
#

oh! good call

frank jungle
#

a better solution is hopefully coming but for now i think its just ending in ? = question

#

or ; in greek i think

icy python
#

nailed it! i updated the prompt to explicitly not use emoji. Just had a conversation with my house 😄

frank jungle
#

there we go

icy python
#

thanks so much for your help I'm really pleased with this 👍