openapi: 3.1.0 info: title: Forethought Triage API description: | The Forethought Triage API classifies customer-support tickets and predicts the most likely category, sub-category, language, sentiment, or other label your Triage model is trained on. Forethought's Implementation Engineering team trains and deploys the model and issues a bearer token. Submit the ticket subject (`text_a`) and body (`text_b`) and the API returns ranked predictions with confidence scores. version: '2025-11-01' contact: name: Forethought Support url: https://support.forethought.ai servers: - url: https://api.forethought.ai description: Production Server security: - BearerAuth: [] tags: - name: Predictions description: Predict labels for customer-support tickets. paths: /api/predict: post: summary: Predict Ticket Triage Labels description: | Submit a ticket subject and body and receive ranked predictions from a Forethought Triage model. Each prediction includes a `prediction_value` (the label) and `prediction_confidence` (0.0 to 1.0). Use the top prediction to route tickets, populate ticket fields, or trigger downstream workflows in Zendesk, Salesforce, or other helpdesks. operationId: predictTriage tags: - Predictions requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/PredictRequest' responses: '200': description: Predictions returned successfully. content: application/json: schema: $ref: '#/components/schemas/PredictResponse' '401': description: Missing or invalid bearer token. '404': description: Triage model not found. '429': description: Request rate limit exceeded. components: securitySchemes: BearerAuth: type: http scheme: bearer bearerFormat: Triage API token schemas: PredictRequest: type: object required: - model_name - text_a properties: model_name: type: string description: The Triage model identifier provided by Forethought. text_a: type: string description: Ticket subject or short title. text_b: type: string description: Ticket body or extended description. PredictResponse: type: object properties: model_name: type: string prediction_confidence: type: number format: float minimum: 0 maximum: 1 top_predictions: type: array items: $ref: '#/components/schemas/Prediction' Prediction: type: object properties: prediction_value: type: string description: The predicted label. prediction_confidence: type: number format: float minimum: 0 maximum: 1