openapi: 3.2.0
info:
  title: assistants Chat API
  version: ''
servers:
- url: https://{tenant}.{region}.qlikcloud.com
  variables:
    region:
      default: us
      description: The region the tenant is hosted in
    tenant:
      default: your-tenant
      description: Name of the tenant that will be called
tags:
- name: Chat
paths:
  /api/v1/assistants/{assistantId}/threads/{threadId}/actions/invoke:
    post:
      tags:
      - Chat
      summary: Execute synchronous prompt
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/promptOutputInvoke'
          headers:
            Etag:
              schema:
                type: string
              description: Assistant version
            Last-Modified:
              schema:
                type: string
              description: Last modified date and time of the Assistant
          description: Prompt is successfully executed.
        '400':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/errorResponse'
          description: The request is in incorrect format
        '403':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/errorResponse'
          description: The user does not have privileges to perform the requested action.
        '404':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/errorResponse'
          description: Assistant is not found.
        '500':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/errorResponse'
          description: Prompt processing error.
      parameters:
      - in: path
        name: assistantId
        schema:
          type: string
          format: uuid
          example: 507f191e810c19729de860ea
        required: true
        description: The ID of the Assistant containing requested Thread
      - in: path
        name: threadId
        schema:
          type: string
          format: uuid
          example: 507f191e810c19729de860ea
        required: true
        description: The ID of the Thread to retrieve
      description: Execute prompt in synchronous non-streaming mode.
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/promptInput'
      x-qlik-visibility: public
      x-qlik-stability: stable
      x-qlik-deprecated: false
      x-qlik-tier:
        tier: '2'
        limit: 100
      operationId: postApiV1AssistantsByAssistantIdThreadsByThreadIdActionsInvoke
      x-operation-id-source: derived
  /api/v1/assistants/{assistantId}/threads/{threadId}/actions/stream:
    post:
      tags:
      - Chat
      summary: Execute asynchronous prompt
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/promptOutput'
          headers:
            Etag:
              schema:
                type: string
              description: Assistant version
            Last-Modified:
              schema:
                type: string
              description: Last modified date and time of the Assistant
          description: Prompt is successfully executed.
        '400':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/errorResponse'
          description: The request is in incorrect format
        '403':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/errorResponse'
          description: The user does not have privileges to perform the requested action.
        '404':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/errorResponse'
          description: Method is not allowed.
        '405':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/errorResponse'
          description: Assistant is not found.
        '500':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/errorResponse'
          description: Prompt processing error.
      parameters:
      - in: path
        name: assistantId
        schema:
          type: string
          format: uuid
          example: 507f191e810c19729de860ea
        required: true
        description: The ID of the Assistant containing requested Thread
      - in: path
        name: threadId
        schema:
          type: string
          format: uuid
          example: 507f191e810c19729de860ea
        required: true
        description: The ID of the Thread to retrieve
      description: Execute prompt in asynchronous streaming mode.
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/promptInput'
      x-qlik-visibility: public
      x-qlik-stability: stable
      x-qlik-deprecated: false
      x-qlik-tier:
        tier: '2'
        limit: 100
      operationId: postApiV1AssistantsByAssistantIdThreadsByThreadIdActionsStream
      x-operation-id-source: derived
components:
  schemas:
    source:
      type: object
      properties:
        chunks:
          type: array
          items:
            type: object
            example:
              text: LLM stands for Large Language Model
              chunkId: 10d347c4-f28a-4faf-93f0-48e781aaf303
          description: List of Chunks used for "AI" generated messages
        source:
          type: string
          example: Reference.md
          description: path to the document used
        documentId:
          type: string
          format: uuid
          example: 10d347c4-f28a-4faf-93f0-48e781aaf303
          description: reference to Document used for "AI" generated messages
        datasourceId:
          type: string
          format: uuid
          example: 10d347c4-f28a-4faf-93f0-48e781aaf303
          description: reference to DataSource used for "AI" generated messages
        knowledgebaseId:
          type: string
          format: uuid
          example: 10d347c4-f28a-4faf-93f0-48e781aaf303
          description: reference to KnowledgeBase used for "AI" generated messages
    promptInput:
      type: object
      properties:
        input:
          type: object
          properties:
            prompt:
              type: string
              example: What is a LLM?
              description: Input prompt string for the Assistant to respond to.
            promptType:
              enum:
              - thread
              type: string
              example: thread
              description: Sets the prompt type to thread.
            includeText:
              type: boolean
              example: true
              description: Returns text from chunks in sources output. Default value is false.
    errorResponseCode:
      type: object
      properties:
        code:
          type: string
          description: 'Custom error codes

            * AE-1 - Internal processing error

            * AE-2 - Incorrect request

            * AE-3 - Authentication issue

            * AE-4 - Prompt is rejected

            * AE-5 - Resource is not found

            * AE-6 - API usage rate limit is exceeded

            * AE-7 - Method is not allowed

            '
        title:
          type: string
        detail:
          type: string
    promptOutput:
      type: object
      properties:
        output:
          type: string
          example: LLM stands for Large Language Model
          description: Assistant's response to the prompt
        sources:
          type: array
          items:
            $ref: '#/components/schemas/source'
          description: List of sources used to generate AI messages
    errorResponse:
      type: object
      properties:
        errors:
          type: array
          items:
            $ref: '#/components/schemas/errorResponseCode'
        traceId:
          type: string
          example: 507f191e810c19729de860ea
    promptOutputInvoke:
      type: object
      properties:
        output:
          type: string
          example: LLM stands for Large Language Model
          description: Assistant's response to the prompt
        sources:
          type: array
          items:
            $ref: '#/components/schemas/source'
          description: List of sources used to generate AI messages
        question:
          type: string
          example: What was the primary goal of the Apollo program?
          description: Question asked by the user for assistant to answer