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

• 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 Azure account with Azure AI service access.

• You have deployed a model in a project.

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.

   

5. Select Details.

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

   

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. Select the Bulk models tab.

4. Click Azure OpenAI endpoints (Chat Completions API).

   tip

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

5. 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.

   

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

   

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.

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.

You can use models from OpenAI-compatible API endpoints as Agent models (in GPT for Excel) and as bulk models. The endpoint requirements are different for each type of model:

• Agent models: The endpoint must be compatible with OpenAI's Responses API and be publicly accessible over the internet through HTTPS. The model must support a context length of at least 32K tokens.

• Bulk models: The endpoint must be compatible with OpenAI's Chat Completions API.

info

Custom API endpoints are available if your space is on the Business or Enterprise plan or pay-as-you-go pricing.

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

• Cloud-based platforms:

  • Anyscale

  • Fireworks AI

  • Together AI

• Local servers:

  • LM Studio

  • LocalAI

  • Ollama

Prerequisites

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

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.

Agent models

Prerequisites

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

• The endpoint supports the /v1/responses OpenAI-compatible API resource path.

• The endpoint is publicly accessible over the internet through HTTPS.

• The model supports a context length of at least 32K tokens.

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

2. In the sidebar, select Custom API endpoints.

3. Select the Agent models tab.

4. Click OpenAI-compatible endpoints (Responses API).

   tip

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

5. Define the endpoint settings:

   • Endpoint URL: Enter the URL of the service that provides the endpoint. The endpoint must support OpenAI's Responses API and be publicly accessible over the internet.

   • Model ID: Enter the ID of the model to use through the endpoint. The model ID is sent to the endpoint in the request payload.

   • Display name (optional): Enter a short name that identifies the model in the Agent model switcher.

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

   

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

You have set up the OpenAI-compatible API endpoint. All space users can now select the model in the Agent model switcher in GPT for Excel.

Bulk models

Prerequisites

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

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

  • /v1/models
  • /v1/chat/completions

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

2. In the sidebar, select Custom API endpoints.

3. Select the Bulk models tab.

4. Click OpenAI-compatible endpoints (Chat Completions API).

   tip

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

5. 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.

   

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

   

   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 an OpenAI-compatible API endpoint and repeat steps 5 and 6.

What's next

• Manage API keys for your space.

• Manage models for your space.

• Manage users in your team.