Skip to main content

Manage API endpoints

As a space owner or admin, you can set up API endpoints for your space. Models available through the endpoints will be accessible to all users in the space

info

API endpoints are available only for Microsoft accounts.

Connect to Azure OpenAI​

GPT for Excel and GPT for Word can connect to Azure AI to use OpenAI models. In Azure AI, each model has a dedicated API endpoint that you use to access the model.

Prerequisites

To connect to an Azure OpenAI API endpoint:

  1. Get the Azure OpenAI API endpoint details.

  2. Set up the Azure OpenAI API endpoint.

Get the Azure OpenAI API endpoint details​

To connect to Azure AI from GPT for Excel and Word, you need the endpoint URL, API key, deployment name, and model name for the Azure OpenAI model you want to use.

  1. Sign in to Microsoft Foundry.

  2. Select the project that has the model you want to use.

  3. Click View deployments.

  4. In Deployed models, in the models list, click the name of the model you want to use.

    Select the deployed model

  5. Select Details.

  6. In the model details, copy the Target URI, Key, Deployment name, and Model name.

    Copy the endpoint details

Set up the Azure OpenAI API endpoint​

  1. Sign in to the GPT for Work dashboard with your Microsoft account.

  2. In the sidebar, select Custom API endpoints.

  3. Click Azure OpenAI API endpoints.

    tip

    If you already have Azure OpenAI API endpoints set up, click Add an Azure OpenAI API endpoint.

  4. Define the endpoint settings:

    • Endpoint URL: Enter the endpoint URL for the model using the Azure OpenAI Chat Completions API endpoint format:

      https://<resource-name>.<domain>/openai/deployments/<deployment-name>/chat/completions?api-version=<api-version>

      If the URL you copied does not use the above format, rewrite it. For example, let's assume that your deployment name is gpt-5.4-mini, the API version is 2025-04-01-preview, and you copied the following URL:

      https://company-resource.services.ai.azure.com/openai/v1/responses

      To rewrite the URL into the Chat Completions API format, replace everything after the domain with /openai/deployments/<deployment-name>/chat/completions?api-version=<api-version> and replace the placeholders with your deployment name and the API version:

      https://company-resource.services.ai.azure.com/openai/deployments/gpt-5.4-mini/chat/completions?api-version=2025-04-01-preview
      info

      The api-version is the Azure OpenAI REST API version, which is not shown on the Microsoft Foundry model details page. As a rule, use 2025-04-01-preview as the value.

    • Model name: Enter the model name.

    • API key: Enter the API key for accessing the endpoint.

    • Display name (optional): Enter a custom display name for the model. If not set, the model switcher uses Model name as the display name.

    Define Azure OpenAI API endpoint settings

  5. Click Check to validate the settings and click Save.

    Save Azure OpenAI API endpoint settings

You have set up the Azure OpenAI API endpoint. All space users can now use the endpoint model in GPT for Excel and GPT for Word.

tip

To set up another Azure OpenAI API endpoint with a different model, click Add an Azure OpenAI API endpoint and repeat the previous two steps.

Set up another Azure OpenAI API endpoint

Connect to OpenAI-compatible API endpoints​

GPT for Excel and GPT for Word can connect to any AI service that provides an OpenAI-compatible API endpoint. You can connect to two main types of services: cloud-based platforms, which provide access to models over the internet with no software installation or setup required on your part, and local servers, which you run on a local machine, such as your own computer or another computer on a local network. Once connected, all models available through the service become available to use in the add-ins.

Here are some popular services that provide an OpenAI-compatible API endpoint:

Prerequisites

  • You are the owner or an admin of the space.

  • Your space is on the Business or Enterprise plan or pay-as-you-go pricing.

  • You have an endpoint URL and, if needed, an API key for accessing the service.

  • The endpoint must support the following OpenAI-compatible API resource paths:

    • /v1/models
    • /v1/completions
info

You cannot currently use models from OpenAI-compatible API endpoints with the Agent.

info

Depending on the model you use, the model settings defined in GPT for Work (temperature (creativity), top-p, frequency penalty, and presence penalty) may not work as expected. If you get bad results, try tweaking the settings.

  1. Sign in to the GPT for Work dashboard with your Microsoft account.

  2. In the sidebar, select Custom API endpoints.

  3. Click OpenAI-compatible endpoints.

    tip

    If you already have OpenAI-compatible API endpoints set up, click Add another OpenAI compatible endpoint.

  4. Define the endpoint settings:

    • Endpoint URL: Enter the URL of the service that provides the endpoint.

    • Display name (optional): Enter a short name that is prefixed to the model names in the model switcher. By default, the model names are prefixed with "custom/".

    • API key (optional): If required, enter the API key for accessing the service.

    Define OpenAI-compatible API endpoint settings

  5. Click Check to validate the settings and click Save.

    Define OpenAI-compatible API endpoint settings
    tip

    If you get an error saying that the URL is invalid, make sure that the endpoint supports:

    • Cross-origin resource sharing (CORS). See the LM Studio setup instructions for macOS or Windows.

    • HTTPS (needed with Safari and with Microsoft Office for Mac). See the LM Studio setup instructions for macOS.

You have set up the OpenAI-compatible API endpoint. All space users can now use all models available through the endpoint in GPT for Excel and GPT for Word.

tip

To set up another OpenAI-compatible API endpoint, click Add another OpenAI compatible endpoint and repeat steps 4 and 5.

Set up another OpenAI-compatible API endpoint

What's next​

On this page