Building better Alexa skills: Voice Model Expression Language (beta)

What is this? This tool (currently beta)This project is work in progress and both functionality and syntax may change without prior notice. is intended to make voice models reusable and easier to maintain. Enrich your Alexa voice model JSON with dynamic 'Spintax' like expressionsExamples
"[help|support] me"
-> 'help me' and 'support me'
"help [ | me]"
-> 'help' and 'help me'
"help [ | me [ | please] ]"
-> 'help' and 'help me' and 'help me please'
and variablesAt the top of the JSON file you may define key/value bindings which can be referred to later in the rules. This makes customization easier in case you have to maintain a number of similar skills for different customers. - which help you to keep configuration separated from the utterance strings. The generator will then process your configuration and create a cleanThe tool will:
- remove duplicates within arrays
- trim and clean up white space
- sort the results
- remove JSON nodes labeled with "description"
file which you can use for your skill configuration in the Alexa Developer Console. You can startThe JSON for the voice model from the Alexa Developer Console build tab will work out of the box. You can add one rule after another to get familiar with what is possible.

You can use this tool to clean up your existing JSONs.
with one rule and grow over time. What do you think? Send any suggestions to

Start conversion
⓵ drop or paste your source file into left editor
⓶ edit/refine your work (e.g. replace 'volvo' with your favourite brand)
⓷ click button labelled with "Start conversion"
⓸ copy result to Alexa Developer Console