Igor is a voice-activated assistant in the same category as Siri and Google Assistant, except it's built as a web app on top of HTML5's x-webkit-speech feature.
<html>
<head>
<script src="http://ajax.googleapis.com/ajax/libs/jquery/1.7.2/jquery.min.js"></script>
<script src="igor.js"></script>
</head>
<body>
<input type="text" id="igor-input" x-webkit-speech />
</body>
</html>You can write your own commands for Igor to use by placing them in a your-project-here.js file in the integrations directory. To set Igor up to match specific words and trigger your functionality, it's as simple as adding a hash to the igor_commands array in igor-commands.js. Example below:
{
'name': 'twitter-status-update',
'matcher': match_all,
'keywords': ['update', 'twitter', 'to'],
'handler': 'update_twitter_status',
'from': 'integrations/twitter.js'
},The previous snippet sets up a command hook that fires whenever "update", "twitter", and "to" are all heard in a user's voice command (since we're using the match_all matcher -- you can also use match_any to match any keyword or write your own custom matchers). When the command hook fires, it calls the update_twitter_status function in integrations/twitter.js, passing in a bit of action context (e.g. the full transcription of what was said, as well as what words matched).
The base Igor script is meant to be as light as possible to prevent causing unneccesary page load times for those who wish to use it. It consists of the following files:
- igor.js - A bootstrap that includes all other neccessary files
- igor-config.js - The configurations available for tweaking Igor
- igor-debug.js - The neccessary functionality required to facilitate debugging Igor, mostly meant for use by developers
- igor-commands.js - The list of commands available to Igor's users, containing what words trigger the command and what function should handle it
- igor-runtime.js - Variables and functionality used to determine which integrations have been initialized at runtime
- igor-initializer.js - The starting point for the end user: this script creates the speech event handler for the #igor-input input field.
To make adding functionality as easy as possible, all 3rd-party functionality is provided in the form of integrations, which are separate scripts that sit in the integrations folder and are injected only if a user tries to use them.
When a command is matched, Igor checks to see if the integration that command's handler comes from has been loaded. If it hasn't, it loads it in. Afterwards, it runs the command's handler, passing along the original text of what was said.