Version 1.4 of the Jovo Framework is now available. Besides new additions to the framework, we have also been busy updating the Jovo Debugger and the Jovo CLI. Read on to learn more.
⭐ Want to support Jovo? Give us a star on GitHub ⭐
To make use of all the newest features, you need the latest version of the jovo-framework package (1.4 or higher) and the jovo-cli (1.2 or higher).
You can globally update the CLI with the following command:
Then, you can either create a new project or update an existing one to have an up-to-date Jovo project:
Let’s take a look at the new features:
An important part of designing and building a voice interaction is the language model. Jovo offers a language model abstraction that can be translated into an Alexa Interaction Model and a Dialogflow Agent for Google Assistant. This is a time saver, but still doesn’t help you in actually designing and testing the model.
Here’s the problem we’ve had while building some more complex voice apps during the last few months: After you have built the language model, how can you make sure that it works correctly? Is there any overlap between intents that could cause mismatches? Is everything still working correctly after I’ve added a few more values? We realized that the practical solution very often is to test the flow manually, over and over.
Manual testing can be a huge pain when you have to go through the whole process of the user interaction over and over again. This why Jovo now offers a language model testing mode that replaces your handler with a very simple one that lets you freely talk to your device and displays the intents and entities that were understood by the voice platform.
To get started, run the model test from the Jovo CLI (again, make sure to update both the CLI and the Framework before you do that):
When you’re now talking to your app on a device, you can say you want, and it returns the following logs:
And even better, the Jovo Debugger got an upgrade with even better model testing capabilities:
Open your debugger and append /model/test to the URL.
Your URL should look like this: https://webhook.jovo.cloud/<your-unique-id>/model/test.
And you can even specify your own tests that create scripts to test if the actual intent is matched:
Let us know, what you think!
We’ve been waiting for this! Earlier this week, Amazon announced the ability for developers to ask for contact information like name, email address, and phone number. Similar to device location, your users can now tap a button in the Alexa companion app and give you the right permissions you need. Previously, people had to use Account Linking (we covered this in a tutorial: Login with Amazon Account Linking) for this, which caused a lot of friction.
We’re happy, that this is now included in the Jovo Framework. Just set up the permissions you need in the Alexa Developer Console, like “Customer Email Address” in the screenshot below:
To ask for the permission, use the following code in your handler, like in this example
You can also find an example file here.
Thanks to our #1 contributor Octavio Menocal, you can now build Alexa Skills for Echo Buttons by using the Alexa Gadgets API. Thanks, Octavio 🙌.
We’re adding a tutorial about this soon. In the meantime, take a look at the Pull Request here.
There are a few more upcoming features we’re working on that will be rolled out in the next few weeks:
Join our Slack community to be among the first to find out: Jovo Community.