Importing Intents

When in the Intents section, you can use the import CTA. import intents CTA

You can import intents using our CSV or our JSON format. You can also import intents from Rasa or Dialogflow ES/CX using our CLI.

HumanFirst JSON format#

Here's an example of intents with training phrases defined in our JSON format. Consult the JSON format documentation for more details and advanced capabilities.

Example#

{
"$schema": "https://docs.humanfirst.ai/hf-json-schema.json",
"intents": [
{
"id": "intent-DIMBHB2C7ZEYZM7FQGKHE3BO",
"name": "card issues"
},
{
"id": "intent-AYXBAIZA4VBXHMAFEYHTAS7N",
"name": "card lost",
"parent_intent_id": "intent-DIMBHB2C7ZEYZM7FQGKHE3BO"
},
{
"id": "intent-XZBR7P4XKZGVROIX3ITOTUVI",
"name": "card stolen",
"parent_intent_id": "intent-DIMBHB2C7ZEYZM7FQGKHE3BO"
},
{
"id": "intent-ZSL5YK3PLFBW7NWL27BC5B3R",
"name": "card damaged",
"parent_intent_id": "intent-DIMBHB2C7ZEYZM7FQGKHE3BO"
}
],
"examples": [
{
"id": "phrase-4553MGLJMZFOLHAEBXCZ36IB",
"text": "where to report lost discover credit card",
"intents": [
{
"intent_id": "intent-AYXBAIZA4VBXHMAFEYHTAS7N"
}
]
}
]
}

CSV format#

Simple 2 column CSV where the 1 column is the utterance/example and the second column is the corresponding intent. For hierarchical intents, you can define the hierarchy with the delimiter of your choice. On import, you can choose Advanced Options and pick the delimiter.

Example#

utterance,intentA
utterance,intentA/intentAB
utterance,intentA/intentAB/intentAC

Rasa#

Rasa yaml#

You can import a Rasa project by providing the Rasa yaml file.

Rasa markdown#

We also support the older Rasa <3.0 markdown format : Rasa markdown format.

Cognigy#

You can import a Cognigy project exported in the Cognigy JSON format.