Skip to main content
The /triage endpoint runs machine learning models to classify support tickets by category and priority, providing confidence scores for each prediction.

​
Endpoint

POST /api/v1/triage

​
Request Body

​
subject
string
required
The subject line of the support ticket to classify
​
body
string
required
The full body/description of the support ticket to classify

​
Response

​
category
string
required
The predicted category for the ticket (e.g., “Billing”, “Technical Support”, “Account Management”)
​
priority
string
required
The predicted priority level for the ticket (e.g., “High”, “Medium”, “Low”)
​
confidence
object
required
Confidence scores for each prediction

​
Example Request

curl -X POST "http://localhost:8000/api/v1/triage" \
  -H "Content-Type: application/json" \
  -d '{
    "subject": "Cannot access my account",
    "body": "I have been trying to log in for the past hour but keep getting an error message. This is urgent as I need to access my dashboard."
  }'

​
Example Response

{
  "category": "Technical Support",
  "priority": "High",
  "confidence": {
    "category": 0.92,
    "priority": 0.87
  }
}

​
Use Cases

The triage endpoint is useful for:
  • Automatic ticket routing - Route tickets to the appropriate support team based on category
  • Priority queuing - Prioritize urgent tickets in the support queue
  • SLA management - Set response time expectations based on priority predictions
  • Analytics - Track common categories and priorities across support tickets

​
Model Training

The ML models used by this endpoint can be trained using the CLI: 
uv run -m src.ml.train
Trained model artifacts are stored in the artifacts/ directory and automatically loaded when the API starts.

​
Confidence Scores

Confidence scores indicate how certain the model is about its predictions:
  • > 0.8 - High confidence, safe to use for automated routing
  • 0.6 - 0.8 - Medium confidence, consider human review
  • < 0.6 - Low confidence, recommend manual classification

Build docs developers (and LLMs) love

Get started for freeTalk to us