Dodeca OpenAI Connector

Overview

The OpenAI connector is a modular extension that allows Dodeca to leverage OpenAI models to analyze and interpret structured reports and data. It also supports MCP (Model Context Protocol) tool sources, enabling models to call external tools and services during a conversation.

Once imported, it enables AI-assisted insights, summarization, and data interpretation directly within the application and views.

Prerequisites

Getting Started

To begin using the Dodeca OpenAI connector, the OpenAI Connector module must be downloaded and imported into the Dodeca tenant. An OpenAI connection must be created and configured to connect to the OpenAI platform.

ℹ️ Note: Supported in Dodeca 8.7.0 and higher, an AI Chat Component module provides a familiar, assistant-ui-based chat component.

OpenAI Connection Properties

Connection Properties

Only a single property is required to configure and test an OpenAI connection:

Property Requirement Description

ApiKey

Required

An OpenAI API key obtained from https://platform.openai.com.

Test Properties

The following properties can be set to execute a prompt and generate a response during testing:

Property Requirement Description

Model

Required

An OpenAI language model.

Effort

Optional

An explicit reasoning effort level, if a reasoning model is chosen.

Instructions

Optional

An instruction set with which to prime the chosen model.

Tool Sources

Optional

MCP server connections that provide tools to the chosen model.

The following success message will be displayed if the connection is successful.

Test Connection

Using the OpenAI Connector

The SAMPLE metadata from Applied OLAP is used in the following example. To add support for and configure the OpenAI Connector, a Dodeca administrator can follow these steps:

  1. Download and import the latest version of the OpenAI Connector and the AI Chat Component modules into a Dodeca tenant.

  2. Restart the Dodeca application. The OpenAI Connector and the AI Chat Component are now available for use.

  3. Create and configure an OpenAI connection using the properties above, then test it to confirm the API key is valid before proceeding.

  4. Go to the Views editor and select an existing Dodeca view you would like to enhance with AI capabilities.

  5. In the View properties grid, find the Components category and press the ellipsis button of the Components property to add a new component.

    A new AI Chat Component can be added and configured via the following properties:

    Property Requirement Description

    ComponentInstanceName

    Required

    The name by which to identify the new component.

    ComponentObjectTypeID

    Required

    The component type to use: AIChatComponent.

    AIConnectorObjectTypeID

    Required

    The AI connector type to use: OpenAIConnector.

    ConnectionID

    Required

    The OpenAI connection to use.

    Model

    Required

    The OpenAI language model to use.

    Effort

    Optional

    The reasoning effort level, if a reasoning model is chosen.

    Instructions

    Optional

    An instruction set with which to prime the chosen model. These can be modified to fit your needs.

    Tool Sources

    Optional

    MCP server connections that provide tools to the chosen model.
    Components Editor

  6. Press OK to keep the new component configuration. The AI Chat Component is now added to your view and ready to test.

  7. Press Preview to test and interact with the new component.

    Test Component

  8. Return to the Views editor and press Commit Selected to save the changes.

Release History

Version 1.1

  • Add support for sessions and multi-turn conversation interaction.

  • Add Effort, Instructions, and Tool Sources properties.

Version 1.0

  • Initial release. Includes support for generating content via the workbook script AIOperations method.