ai-austin / bardvoice

Voice activated Python interface for Bard AI. Implements open sourced reverse engineered Bard API, local text to speech and OpenAI Whisper for local transcribing.

Google Bard Voice Activated Assistant

Google AI BARD powered Python voice assistant. Implements a free Bard API and OpenAI Whisper for local transcribing.

YouTube tutorial and preview

https://youtu.be/vfbmlRSgj9Q

NOTE: Open AI whisper will not run on Python versions newer than 3.10. Python 3.11 and newer will not run this program.

Authentication

Sign into your account with Bard access at https://bard.google.com/

• F12 for developer console
• Copy the values
  • Session: Go to Application → Cookies → __Secure-1PSID and __Secure-1PSIDTS. Copy the values of those.
  • Paste the values in line 10 & 11 for the token and ts_token variables inside main.py

Installation

Clone GitHub repo.

git clone https://github.com/Ai-Austin/BardVoice

Change directory to bard_voice.

cd bard_voice

Paste your Bard Token into line 10 of the main.py file, replacing {YOUR BARD TOKEN}.

token = "{YOUR BARD TOKEN}"

Install Python dependencies (use pip3 if your system requires).

pip install -r requirements.txt

Whisper requires the command-line tool ffmpeg to be installed on your system, which is available from most package managers:

# on Ubuntu or Debian
sudo apt update && sudo apt install ffmpeg

# on Arch Linux
sudo pacman -S ffmpeg

# on MacOS using Homebrew (https://brew.sh/)
brew install ffmpeg

# on Windows using Chocolatey (https://chocolatey.org/)
choco install ffmpeg

# on Windows using Scoop (https://scoop.sh/)
scoop install ffmpeg

Startup Voice Assistant

python main.py

bardvoice's Issues

error transcribing audio

I've named the mp3 file but it didn't even detect or write the sound i dont'n know how to fix it?
PLeasseeee heeeeelllpppp

no sound

I get no sound back it creates the audio file for wake detected and that's it I am on a Mac m2 running it on conda 3.10 env

Say "google" to wake me up.

/Users/user/miniforge3/envs/bard/lib/python3.10/site-packages/whisper/transcribe.py:114: UserWarning: FP16 is not supported on CPU; using FP32 instead
warnings.warn("FP16 is not supported on CPU; using FP32 instead")
No wake word found. Try again.

Say "google" to wake me up.

uncertain - Bard not found

normally i have been able to follow along and everything works - even tho on windows- this one tho it is unable to find the Bard module despite being pip installed.

pip installed requirements.txt added in the token - but no dice there - tried manually pip installing Bard but same response.

thank you as always for your tutorials and your code examples. They have helped me learn and develop.

---

I was able to sort above my manually installing every package not sure what is going on with the requirements.txt file not installing

but now

asked GPT 4 it told me it was potentially an error with the token - so i went and made sure i got the right one form the dev console - still having an error tho and not sure how to solve

sorry for the newbishness.

correction for PyAudio==2.3.0

whe you run the pip install -r requirements.txt you will get an error in the terminal because of a typo in readme particularly for PyAudio==2.3.0. i noticed it right away and just thought i would bring it up to be updated. anyway thanks for this code man, you are doing awesome

retrieval of access token securely and automatically

I havent thought about it yet, but im wondering how you can just retrieve your access token instead of the F12 manual method