Skip to main content

XTM Cloud 26.2

OpenAI GPT LLM configuration

You can connect your OpenAI GPT model to XTM Cloud either via API or Azure OpenAI. For details, see the table below.

To check which OpenAI GPT models are currently supported via API and Azure OpenAI, see the Supported models section in https://xtm.ai.transifex.com/docs#section/LLM-Configuration.

Procedure. To configure your OpenAI GPT LLM connection in XTM Cloud:

  1. Select Configuration > Settings > AI Control Center.

  2. Select the LLM Configuration tab.

  3. In the LLM Configuration section, find the OpenAI GPT section.

  4. In the OpenAI GPT section, fill in the relevant fields, as described below:

    Important

    To make the settings that relate to the use of OpenAI GPT as an MT engine, select Configuration > Settings > Translation > Machine translation tab. For details, see Connecting XTM Cloud to your OpenAI GPT engine.

    Table 49. OpenAI GPT – configuration settings

    Setting

    Description

    Mandatory/Optional

    Type

    Valid values

    Connection method

    Use this setting to specify how your XTM Cloud instance is to connect to your OpenAI GPT model.

    Project Managers will be able to change this setting for the relevant customer.

    Mandatory

    Radio buttons

    • OpenAI GPT.

      If you select this option, the API Key setting is displayed.

    • Azure OpenAI.

      If you select this option, these settings are displayed:

      • Azure API key.

      • Endpoint.

      • Deployment names.

      • Select default deployment name.

    API Key

    This setting is displayed if you have selected OpenAI GPT as your Connection method.

    API Key for OpenAI GPT.

    Project Managers will be able to change this setting for the relevant customer.

    Mandatory if you have selected OpenAI as your Connection method.

    Text field

    Alphanumeric

    Azure API key

    This setting is displayed if you have selected Azure OpenAI as your Connection method.

    Azure API Key for OpenAI GPT service.

    Project Managers will be able to change this setting for the relevant customer.

    Mandatory if you have selected Azure OpenAI as your Connection method.

    Text field

    Alphanumeric

    Endpoint

    This setting is displayed if you have selected Azure OpenAI as your Connection method.

    URL for OpenAI GPT.

    Project Managers will be able to change this setting for the relevant customer.

    Mandatory if you have selected Azure OpenAI as your Connection method.

    Text field

    Alphanumeric

    Deployment names

    This setting is displayed if you have selected Azure OpenAI as your Connection method.

    Enter one or more deployment names for the OpenAI GPT service that have been specified during the deployment of a model, in Azure OpenAI Studio.

    If you are using the AI Translate & TQI feature, two deployment names are required to obtain the best-quality results. Even if you use two LLMs (such as GPT and Gemini) to run the feature. We recommend that you use one deployment name for a small model (for example, 4o-mini) and one for a large model (for example, 4o).

    Project Managers will be able to change this setting for the relevant customer.

    Mandatory if you have selected Azure OpenAI as your Connection method.

    Text field

    Alphanumeric

    Select default deployment name

    This setting is displayed if you have selected Azure OpenAI as your Connection method.

    Out of all Deployment names select one that will be the system default.

    Project Managers will be able to change this setting for the relevant customer.

    Mandatory if you have selected Azure OpenAI as your Connection method.

    Dropdown

    Specified in the Deployment names setting.


  5. At the bottom of the page, select Save.

Result: OpenAI GPT LLM is connected to your XTM Cloud instance.

You can make the OpenAI GPT settings related to its use as an MT engine in Configuration > Settings > Translation > Machine translation tab. For details, see Connecting XTM Cloud to your OpenAI GPT engine.

Project Managers can enable or disable OpenAI GPT at customer level. They can also select a different connection method and/or connection details for a particular customer.

At project level, Project Managers can add OpenAI GPT to a new project while creating it, or to an existing one in the Project Editor. Additionally, they can enable or disable it for a particular language pair. Segments processed by OpenAI GPT are counted as MT matches in the project Metrics and Statistics, in the Project Editor. Project Managers can also enable or disable OpenAI GPT for a particular project template(s).

In XTM Workbench, OpenAI GPT MT matches are labeled as OpenAI GPT MT.

In this section: