Docker Speak Windows 11 solution

[BrokenStandards]

Getting speak mode to work in docker can be a bit challenging. Not only is the project missing some things necessary for speak mode to work in the dockerfile and requirements, But getting sound from docker in Windows is already a task to deal with itself! I will provide a solution here.

Add necessary packages to the Dockerfile and requirements

Add the following packages to your Dockerfile and requirements file:

# Speak mode
# Note we do not install python3-gi here. Last I tested it does not even see it.
RUN apt-get update \
    && apt -y install libcairo2-dev libgirepository1.0-dev python3-gst-1.0 gir1.2-gstreamer-1.0 gstreamer1.0-plugins-good gstreamer1.0-plugins-bad gstreamer1.0-plugins-base

Add the following line somewhere near the top of your requirements.txt file:

PyGObject

If you are not on windows, this is probably enough (and if it is not skip to the last step and adjust the relevant pulse variables).

Install Pulseaudio Server on windows

The first step is to set up a PulseAudio server outside of the Docker container. In this guide, we will install it in an Ubuntu WSL instance. Note you could instead install pulseaudio server on windows itself, but we will not cover that here. If you haven't already installed WSL and Ubuntu, please do so before proceeding. Once you have Ubuntu installed, follow these steps to install PulseAudio:

1. Open the terminal in your Ubuntu WSL instance.
2. Run the following command to install PulseAudio:

sudo apt update
sudo apt install pulseaudio

Follow the prompts to complete the installation.

Now that you have set up the PulseAudio server on your Windows system, let's proceed to update the Docker Compose file.

Update the Docker Compose file

To enable sound in the Docker container, you need to make two modifications to your Docker Compose file:

1. Add the PULSE_SERVER environment variable: Open your Docker Compose file and locate the auto-gpt section. Add the following environment variable:

environment:
  - PULSE_SERVER=/mnt/wslg/PulseServer

This environment variable specifies the path to the PulseAudio server socket within the container.

2. Mount the wslg volume: The wslg volume allows the Docker container to access the PulseAudio socket in the WSL instance. Add the following volume definition to the auto-gpt section in your Docker Compose file:

    volumes:
      - \\wsl$\Ubuntu\mnt\wslg:/mnt/wslg

Save the changes to your Docker Compose file.

With these modifications, your Docker container will be able to access the PulseAudio server running in the WSL instance.

[QNimbus]

This worked for me, with the addition that my volume mount in the docker-compose.yml file had to be changed to this:

...
    volumes:
      - \\wsl.localhost\Ubuntu-22.04\mnt\wslg:/mnt/wslg
...

[Ahmedfq]

Thanks for creating this thread and detailing how to fix the issue. I followed the steps but still get the following error:

"Exception in thread Thread-1 (speak):
Traceback (most recent call last):
File "/usr/local/lib/python3.10/threading.py", line 1016, in _bootstrap_inner
self.run()
File "/usr/local/lib/python3.10/threading.py", line 953, in run
self._target(*self._args, **self._kwargs)
File "/app/autogpt/speech/say.py", line 27, in speak
success = voice_engine.say(text, voice_index)
File "/app/autogpt/speech/base.py", line 45, in say
return self._speech(text, voice_index)
File "/app/autogpt/speech/gtts.py", line 21, in _speech
playsound("speech.mp3", True)
File "/usr/local/lib/python3.10/site-packages/playsound.py", line 91, in _playsoundNix
import gi
ModuleNotFoundError: No module named 'gi'"

I probably missed something but cannot seem to identify what. Can someone please help, thanks.

[Ahmedfq]

Oh, by the way I am trying to use elevenlabs for TTS and my .env file looks like this

// TEXT_TO_SPEECH_PROVIDER - Which Text to Speech provider to use (Default: gtts)
TEXT_TO_SPEECH_PROVIDER=gtts

// Only if TEXT_TO_SPEECH_PROVIDER=streamelements
// STREAMELEMENTS_VOICE - Voice to use for StreamElements (Default: Brian)
// STREAMELEMENTS_VOICE=Brian

// Only if TEXT_TO_SPEECH_PROVIDER=elevenlabs
// ELEVENLABS_API_KEY - Eleven Labs API key (Default: None)
ELEVENLABS_API_KEY=<<My_API_Key>>

// ELEVENLABS_VOICE_ID - Eleven Labs voice ID (Example: None)
ELEVENLABS_VOICE_ID=<<My_Voice_ID>>

[Metochoi]

Im in the latest version of AutoGPT using Docker and there is no requirements.txt. is there another way to fix TTS for the latest version?