Jovo Model
Cross-platform language model converter. Define your intents, entities, and training phrases once — Jovo Model converts them to Alexa Interaction Model, Dialogflow Agent, Rasa NLU, NLP.js, and more.
Supported Platforms
- ✓ Amazon Alexa (ASK)
- ✓ Google Dialogflow
- ✓ Rasa NLU
- ✓ NLP.js
- ✓ Microsoft LUIS
Key Features
- • Single-source language model
- • CLI integration (
jovo build) - • Custom entity types
- • Multi-locale support
Installation
Included by default in all Jovo v4 projects.
How It Works
The Jovo Model is a JSON format that describes your app's language model in a platform-agnostic way. You define intents with training phrases, entities with synonyms, and input types — then the jovo build command converts this into the native format for each platform you're targeting.
Instead of maintaining separate Alexa Interaction Models and Dialogflow agent configurations, you maintain one models/en.json file. When you add a new intent or training phrase, it automatically propagates to all platforms.
Example Model
{
"invocation": "my voice app",
"intents": {
"OrderPizzaIntent": {
"phrases": [
"order a {size} {topping} pizza",
"I want a {size} pizza with {topping}",
"get me a pizza"
],
"entities": {
"size": { "type": "PizzaSize" },
"topping": { "type": "PizzaTopping" }
}
}
},
"entityTypes": {
"PizzaSize": { "values": [{ "value": "small" }, { "value": "medium" }, { "value": "large" }] }
}
}