Skip to content

uberduck-ai/openduck

Repository files navigation

Openduck

Goals

Building an open source talking plush with a voice you can talk to.

Our kids should grow up in a household full of hackable droids they can play with, work on, and learn from.

Setup

  1. Copy the .env.example to .env and fill in real values.

  2. Download server-side models and put them inside of openduck-py/models (TODO: share models.)

Install Dependencies

  1. Install espeak (brew install espeak on Mac OS or sudo apt-get install espeak-ng on Debian Linux).

    • There may be other environment variables you need to set on Mac OS. I had to: export PHONEMIZER_ESPEAK_LIBRARY=/opt/homebrew/Cellar/espeak/1.48.04_1/lib/libespeak.dylib.
  2. pip install -r openduck-py/requirements.txt

Without Docker

cd openduck-py
uvicorn openduck_py.routers.main:app --reload --env-file ../.env

With Docker

docker-compose up

Run the client

Simple Python Client

cd clients/simple
# Lighter-weight requirements
pip install -r requirements.txt
python simple_bot.py --record

Conversation Review

This will start up streamlit on port 8501 so make sure that port is forwarded if you are runnin on ssh.

Install streamlit

pip install streamlit

Run streamlit

cd openduck-py
streamlit run observability.py

Alternative is to use Dockerfile

Troubleshooting

  • The Simple Python client sometimes has strange audio playback bugs. You can try restarting your OS's audio services, e.g. sudo pkill coreaudiod on Mac OS.