ESP + Microphone to control HA Assist with external Speaker. Problem with define the voice

hi guyes i add an esp with microphone to talk to HA Assistant. And the output from the Assistant is send over integrated media.player speaker. In my Case i am using Alexa echo dot for the voice answers from HA Assist.

My problem is that alexa is using her own voice. I want to give the output with the voices from HA. In HA i tested the code in the services section. And it working without problems. But when i add the Code in the ESP then i get an Error Message.

This is the code which worked in services

service: tts.cloud_say
data:
  cache: false
  entity_id: media_player.echo_wohnzimmer
  message: >-
    Hello this code here is working
  language: de-DE
  options:
    voice: ChristophNeural

This code here working too in the service section

service: tts.speak
data:
  media_player_entity_id: media_player.echo_wohnzimmer
  message: Dies ist eine Testnachricht
  entity_id: tts.home_assistant_cloud
  language: de-DE
  options:
    voice: MajaNeural

and by ESP i got this error message:
on_tts_start:

homeassistant.service: 
      service: tts.cloud_say
      data:

Must be string, got <class 'esphome.helpers.OrderedDict'>. did you forget putting quotes around the value?.

options: 
          voice: ChristophNeural
        entity_id: media_player.echo_wohnzimmer
        message: !lambda |-
          return x;

i think ESP want here a different code format but i have no idea how…
can someone help please?

what mean this?
Must be string, got <class ‘esphome.helpers.OrderedDict’>. did you forget putting quotes around the value?.

What I can think of is this:

The working sample uses message:>- as notation and then the message on the next line. Like with templates no quotes are required here
However, the error version does use a one line message. For templates quotes are required then. so try:

message: "Dies ist eine Testnachricht"

or

message: >-
   Dies ist eine Testnachricht

I found the solution.
The ESPHome can not use the options parameter. For that you must create a date_template. Then it works. Now i can change the voice by alexa