Twitch Conversation

An asynchronous NodeJS app connecting IBM Watson's Conversation and Twitch

Features

Listens to messages in Twitch chat channels and feeds them to the Conversation service.

Replies with the Conversation response through a connected bot.

Connects to MongoDB to store conversation context for each user.

Installation

Download the repository and then type "npm install" while in the 'src/' directory.

Setup

Create and run a MongoDB server. Then create a new database with a new collection.

Create a 'config.json' file in the 'src/' directory. Follow the format below:

{
    "twitch": {
        "username": "(Bot's Twitch account username here)",
        "channel": "(Twitch channel for bot to join, prefixed with hastag aka. pound aka. # symbol)",
        "password": "(Bot's Twitch OAuth password here. Find it at http://www.twitchapps.com/tmi/ )"
    },
    "conversation": {
        "username": "(Conversation username here)",
        "password": "(Conversation password)",
        "workspace_id": "(Conversation workspace_id)",
        "url": "(Conversation url)"
    },
    "mongodb": {
        "url": "mongodb://(Address to MongoDB server here, localhost if on same computer as app):(Port, default is 27017)/(MongoDB database name here)",
        "collection": "(MongoDB collection name here)"
    }
}

Usage

You can now start the app by typing "npm start" while in the 'src/' directory.

The bot responds to Twitch messages that contain the Twitch username of the bot.