> ## Documentation Index
> Fetch the complete documentation index at: https://docs.enneo.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Evaluate an AI agent prompt

> Run the LLM prompt that belongs to the given AI agent against a sample input and return the
raw model output. Used by the prompt editor in ops-fe to iterate on prompt wording without
deploying the agent.




## OpenAPI

````yaml https://dev.enneo.dev/api/mind/docs/open-api post /aiAgent/{id}/prompt-eval
openapi: 3.0.0
info:
  version: '1'
  title: enneo.MIND API
  description: This describes the API of enneo Mind, the main ticketing backend
  contact:
    name: enneo GmbH
    email: richard@enneo.ai
  license:
    name: Proprietary software
    url: https://enneo.ai
servers:
  - url: https://demo.enneo.ai/api/mind
    description: Production server, demo client
  - url: https://main.enneo.dev/api/mind
    description: Development main branch
  - url: http://localhost:8005/api/mind
    description: Local development server
security:
  - bearerAuth:
      - api
  - cookieAuth:
      - api
paths:
  /aiAgent/{id}/prompt-eval:
    post:
      tags:
        - AI Agents
      summary: Evaluate an AI agent prompt
      description: >
        Run the LLM prompt that belongs to the given AI agent against a sample
        input and return the

        raw model output. Used by the prompt editor in ops-fe to iterate on
        prompt wording without

        deploying the agent.
      operationId: aiAgentPromptEval
      parameters:
        - name: id
          in: path
          required: true
          description: ID of the AI agent.
          schema:
            type: integer
          example: 42
      requestBody:
        required: true
        content:
          application/json:
            schema:
              type: object
              properties:
                prompt:
                  type: string
                  description: >-
                    Prompt template to evaluate (overrides the agent's stored
                    prompt for this call).
                ticketId:
                  type: integer
                  description: Optional ticket id to inject ticket context into the prompt.
                variables:
                  type: object
                  description: Additional template variables.
      responses:
        '200':
          description: Successful operation
          content:
            application/json:
              schema:
                type: object
                properties:
                  output:
                    type: string
                    description: Raw LLM output.
                  tokens:
                    type: object
                    description: Token usage breakdown.
        '400':
          description: Invalid input
        '403':
          description: Unauthorized
components:
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      bearerFormat: JWT
      description: JWT-based authentication
      x-scopes:
        api: Full access to the API
    cookieAuth:
      type: apiKey
      in: cookie
      name: connect.sid
      description: Cookie-based authentication
      x-scopes:
        api: Full access to the API

````