AI Writing Tool for Live Performance

About

This was developed for the NYC Media Lab and Bertelsmann Creative Industries and AI Challenge, Fall 2023. I lead a fabulous team with Patrick Coleman, Ash Smith, Ryan Schmaltz, and Jinku Kim. Our project employs AI for pre-preproduction (co-creation and world building) and performance (scripts as live artifacts, real-time AI).

See https://roberttwomey.com/2023/12/nyc-media-lab-x-bertelsmann-ai-and-creative-industry-challenge/

Read more here: https://engineering.nyu.edu/research-innovation/entrepreneurship/nyc-media-lab/projects/creative-industries-and-ai-challenge

Setup

Install Node and NPM

update package list:

sudo apt update

install node and npm:

sudo apt install nodejs npm

follow onscreen prompts to restart daemons

EDIT: that didn't work. do this instead

remove old install

sudo apt purge nodejs npm

install nvm

curl -o- https://raw.githubusercontent.com/creationix/nvm/v0.33.8/install.sh | bash

source nvm bashrc

export NVM_DIR="$HOME/.nvm"
[ -s "$NVM_DIR/nvm.sh" ] && \. "$NVM_DIR/nvm.sh"  # This loads nvm
[ -s "$NVM_DIR/bash_completion" ] && \. "$NVM_DIR/bash_completion"  # This loads nvm bash_completion

install new node via nvm

nvm install node

this gets us a current version (v21.4.0)

reference: from here https://shandou.medium.com/upgrade-nodejs-on-ubuntu-amazon-ec2-e686646e5725

Install our Code

clone it

git clone https://github.com/roberttwomey/radio-writing-tool/

cd radio-writing-tool

Checkout demo branch

git checkout medialab-demo

Install all our dependencies

npm install

Set up environment

in the radio-writing-tool/ folder, create .env file OpenAI Key info. Copy this from your local.

touch .env

In the file, copy the following text. Replace XXXXXX with your openai api key (copied from your openai account)

OPENAI_API_KEY=XXXXXX

echo "OPENAI_API_KEY=XXXXXX" > .env

Usage

Run with node (standalone) or with pm2 (kind of as a service)

node writing-tool.js

or

pm2 start writing-tool.js

Visit in Browser

Open localhost:8080 in Browser (on mac, Safari or Chrome best implement webSpeech). (speechRec isn't working in edge)

Running with pm2

Start the app:

pm2 start writing-tool.js

Inspect the log:

pm2 log writing-tool.js

Stop the writing tool:

pm2 stop writing-tool.js

Restart (if you made changes):

pm2 restart writing-tool.js

read logs

logs are in

~/.pm2/logs

References

• Deploying p5 sketch with node: https://github.com/processing/p5.js/wiki/p5.js,-node.js,-socket.io
• PM2 to deploy a sketch: https://pm2.keymetrics.io/