Anybody have a working config for a
ESP32 Development Board CP2102 with a
INMP441 Omnidirectional Microphone?
Just playing around and don’t have a speaker on it yet. Really struggling to get any of the findings posted here working though.
Figured somebody must have one worrying right???
Thanks for any help.
See SpotPear "DeepSeek Voice Chat" config and GitHub - RealDeco/xiaozhi-esphome: Alternative code to use xiaozhi ai devices in esphome/home assistant., those devices are esp32-s3+INMP441+max98357a+display, just adapt/ strip down the example to your needs (don’t expect the quality to be sufficient, though).
Between Gemini and grok I came up with this. Seems to work so far. 
Blue light turns on when it hears the wake word and blinks when listening for my voice request.
substitutions:
voice_assist_idle_phase_id: '1'
voice_assist_listening_phase_id: '2'
voice_assist_thinking_phase_id: '3'
voice_assist_replying_phase_id: '4'
voice_assist_not_ready_phase_id: '10'
voice_assist_error_phase_id: '11'
voice_assist_muted_phase_id: '12'
esphome:
name: voice-box
friendly_name: voice
platformio_options:
board_build.flash_mode: dio
on_boot:
priority: 600
then:
- delay: 30s
- if:
condition:
lambda: return id(init_in_progress);
then:
- lambda: id(init_in_progress) = false;
# Boot LED Test
- logger.log: "Built-in LED Boot Test: ON"
- light.turn_on:
id: va_led
brightness: 1.0
- delay: 1s
- logger.log: "Built-in LED Boot Test: OFF"
- light.turn_off:
id: va_led
- delay: 1s
- logger.log: "Built-in LED Boot Test Complete."
esp32:
board: esp32dev
#variant: esp32s3
framework:
type: esp-idf
version: recommended
components:
# Enable logging
logger:
level: DEBUG
# Enable Home Assistant API
api:
encryption:
key: "" # REPLACE with your own key
# Enable OTA updates
ota:
- platform: esphome
password: "" # REPLACE with your own password
wifi:
ssid: "" # REPLACE with your SSID
password: "" # REPLACE with your password
# Enable fallback hotspot (captive portal) in case wifi connection fails
ap:
ssid: "Esp32-Mic-Speaker"
password: "9vYvAFzzPjuc"
captive_portal:
i2s_audio:
- id: i2s_in
i2s_lrclk_pin: GPIO33 ##INMP441-WS
i2s_bclk_pin: GPIO18 ##INMP441-SCK
- id: i2s_out
i2s_lrclk_pin: GPIO16 ##Max98357 - LRC
i2s_bclk_pin: GPIO20 ###Max98357 - BCLK
## l/R pin on ##INMP441 is connected to ground
microphone:
platform: i2s_audio
id: external_microphone
adc_type: external
i2s_audio_id: i2s_in
i2s_din_pin: GPIO13
channel: left
pdm: false
speaker:
platform: i2s_audio
id: external_speaker
dac_type: external
i2s_audio_id: i2s_out
i2s_dout_pin: GPIO14 ###Max98357 - DIN
#mode: mono
output:
- platform: ledc
pin: GPIO2
id: builtin_led_output
frequency: 1000Hz
inverted: true
light:
- platform: monochromatic
name: "Built-in LED"
id: va_led
output: builtin_led_output
default_transition_length: 0.2s
restore_mode: ALWAYS_OFF
voice_assistant:
id: va
microphone: external_microphone
speaker: external_speaker
use_wake_word: true
noise_suppression_level: 4
auto_gain: 26dBFS
volume_multiplier: 1.5
on_wake_word_detected:
- lambda: |-
id(voice_assistant_phase) = ${voice_assist_listening_phase_id};
ESP_LOGD("custom", "on_wake_word_detected - Phase: %d, turning on LED for wake word", id(voice_assistant_phase));
id(va_led).turn_on().set_brightness(1.0).perform();
on_stt_end:
then:
- logger.log:
format: "==> Received voice command: %s"
args: [ 'x.c_str()' ]
on_listening:
- lambda: |-
ESP_LOGD("custom", "on_listening triggered - Phase: %d, starting blink for listening", id(voice_assistant_phase));
- script.execute: blink_listening
on_stt_vad_end:
- lambda: |-
id(voice_assistant_phase) = ${voice_assist_thinking_phase_id};
ESP_LOGD("custom", "on_stt_vad_end triggered - Phase: %d, setting LED dim (waiting to speak)", id(voice_assistant_phase));
id(va_led).turn_on().set_brightness(0.5).perform();
on_tts_stream_start:
- lambda: |-
id(voice_assistant_phase) = ${voice_assist_replying_phase_id};
ESP_LOGD("custom", "on_tts_stream_start triggered - Phase: %d, setting LED bright (replying)", id(voice_assistant_phase));
- light.turn_on:
id: va_led
brightness: 1.0
on_tts_stream_end:
- lambda: |-
id(voice_assistant_phase) = ${voice_assist_idle_phase_id};
ESP_LOGD("custom", "on_tts_stream_end triggered - Phase: %d, turning off LED", id(voice_assistant_phase));
id(va_led).turn_off();
on_error:
- if:
condition:
lambda: return !id(init_in_progress);
then:
- lambda: |-
id(voice_assistant_phase) = ${voice_assist_error_phase_id};
ESP_LOGD("custom", "on_error triggered - Phase: %d, starting flash_error script", id(voice_assistant_phase));
- script.execute: flash_error
- delay: 1s
- if:
condition:
switch.is_on: use_wake_word
then:
- lambda: |-
id(voice_assistant_phase) = ${voice_assist_idle_phase_id};
id(va_led).turn_off();
else:
- lambda: |-
id(voice_assistant_phase) = ${voice_assist_muted_phase_id};
id(va_led).turn_on().set_brightness(0.3).perform();
on_client_connected:
- if:
condition:
switch.is_on: use_wake_word
then:
- voice_assistant.start_continuous
- lambda: |-
id(voice_assistant_phase) = ${voice_assist_idle_phase_id};
id(va_led).turn_off();
ESP_LOGD("custom", "on_client_connected wake word branch - Phase: %d, LED off", id(voice_assistant_phase));
else:
- lambda: |-
id(voice_assistant_phase) = ${voice_assist_muted_phase_id};
ESP_LOGD("custom", "on_client_connected muted branch - Phase: %d, setting LED dim (muted)", id(voice_assistant_phase));
id(va_led).turn_on().set_brightness(0.3).perform();
on_client_disconnected:
- lambda: |-
id(voice_assistant_phase) = ${voice_assist_not_ready_phase_id};
ESP_LOGD("custom", "on_client_disconnected triggered - Phase: %d, starting flash_disconnected script", id(voice_assistant_phase));
- script.execute: flash_disconnected
switch:
- platform: template
name: Use Wake Word
id: use_wake_word
optimistic: true
restore_mode: RESTORE_DEFAULT_ON
on_turn_on:
- if:
condition:
lambda: return !id(init_in_progress);
then:
- lambda: |-
id(voice_assistant_phase) = ${voice_assist_idle_phase_id};
id(va_led).turn_off();
ESP_LOGD("custom", "switch on_turn_on - Phase: %d, LED off", id(voice_assistant_phase));
- if:
condition:
not:
- voice_assistant.is_running
then:
- voice_assistant.start_continuous
on_turn_off:
- if:
condition:
lambda: return !id(init_in_progress);
then:
- voice_assistant.stop
- lambda: |-
id(voice_assistant_phase) = ${voice_assist_muted_phase_id};
id(va_led).turn_on().set_brightness(0.3).perform();
ESP_LOGD("custom", "switch on_turn_off - Phase: %d, setting LED dim (muted)", id(voice_assistant_phase));
globals:
- id: init_in_progress
type: bool
restore_value: no
initial_value: 'true'
- id: voice_assistant_phase
type: int
restore_value: no
initial_value: ${voice_assist_not_ready_phase_id}
# ====================================================================
# SCRIPTS FOR NON-BLOCKING FLASHING AND BLINKING
# ====================================================================
script:
- id: flash_error
mode: restart
then:
- repeat:
count: 3
then:
- light.turn_on:
id: va_led
brightness: 1.0
- delay: 200ms
- light.turn_off:
id: va_led
- delay: 200ms
- id: flash_disconnected
mode: restart
then:
- repeat:
count: 5
then:
- light.turn_on:
id: va_led
brightness: 1.0
- delay: 100ms
- light.turn_off:
id: va_led
- delay: 100ms
- id: blink_listening
mode: restart
then:
- repeat:
count: 10
then:
- light.turn_on:
id: va_led
brightness: 1.0
- delay: 200ms
- light.turn_off:
id: va_led
- delay: 200ms
For voice to work well you will need at least some psram. Preferably a n16r8 board. As far as I can tell this board does not have that.
And the best place for working code is the examples in this thread.
Does the sound work? I’m still practicing, I found your code, but the trouble is, the sound doesn’t even crack.
That is probably due to the board you are trying to use. You may be better off basing your code on this.
wake-word-voice-assistants/m5stack-atom-echo at main · esphome/wake-word-voice-assistants · GitHub
My yaml works fine on the esp32 n16r8, sound plays perfectly.
Unfortunately, I only have esp-wroom-32 available + inmp441 + MAX98357A
In my experience you will have audio issues unless you use the N16R8 boards. They tend to crash if using esp32 without psram.
Thank you. For the advice.