Generates a draft reply from a message, email, ticket or conversation, taking the provided context into account.
It is useful for supporting communication flows where coherent, relevant replies adapted to the channel are needed.
Overview
Endpoint
- POST /api/v1/text/reply
Use cases
- Suggest replies for customer care operators
- Support ticket management
- Partially automate communication workflows
- Integrate reply generation into application flows via API
Technical specifications
- INPUT: prompt | content
- OUTPUT: content
- ACCESS: API REST
- CHANNELS: Cloud Management Platform, API Platform
Implementation examples
Input
curl --location 'https://api.ai.devops.aruba.it/api/services/textmanagement/v1/text/reply' \
--header 'Content-Type: application/json' \
--header 'ai-platform-key: ••••••' \
--header 'conversation-id: ••••••' \
--header 'Authorization: ••••••' \
--data '{
"prompt": "accetta l'\''invito",
"content": "Ciao, sei libero domani alle 18? Vorrei incontrarti al bar vicino al parco per parlare del progetto."
}'
Output
{
"content": "Ciao! Sì, domani alle 18 va benissimo. Ci vediamo al bar vicino al parco. A presto!",
"ai_platform_usage": {
"service": "text-management",
"namespace": "ai-platform",
"prompt_tokens": 428,
"completion_tokens": 137,
"total_tokens": 565
},
"ai_generated": true
}
Cost
0,0318 €
+ VAT / 1M token (input)
0,1533 €
+ VAT / 1M token (output)
NOTE
Submitted data is not used to train the models.