Even though you’ll be running your skill’s fulfillment code locally during development, the interaction model still needs to be deployed so that Alexa can understand user utterances, map them to intents, and process any custom types and utterances you may have defined.
For an Alexa-hosted skill, deploying a skill means committing and pushing the code to the Git repository:
| $ git commit -am "...commit message..." |
| $ git push |
On the other hand, if you have chosen explicit AWS Lambda hosting for your skill’s fulfillment code, you’ll need to use the ASK CLI to deploy the skill:
| $ ask deploy |
This will deploy the entire skill, including the fulfillment code. If you only want to deploy the skill metadata, which includes the interaction model, you can specify “skill-metadata” as the target:
| $ ask deploy --target skill-metadata |
Either way, this ensures that the interaction model is available in the Alexa platform. Once the skill’s interaction model has been deployed, you are ready to start running the skill locally.