AI API Access

IS&T can provide faculty, staff, researchers, and departments with access to Large Language Model (LLM) API keys for use in academic and research tasks. If you are interested in LLM access for administrative tasks, please contact us first to discuss resources that might already be available. This service offers keys for popular AI models through Microsoft Azure OpenAI and Amazon Bedrock, enabling the BU community to leverage cutting-edge generative AI capabilities in a secure, compliant, and convenient manner.

Benefits

• Strong Data Protection: The service ensures your data remains confidential and secure under BU’s existing vendor agreements. Any prompts or results you send are not shared externally or used to train the providers’ models. All LLMs offered through this service are authorized for data up to Confidential.

• Centralized Billing through IS&T: All API usage costs are consolidated on a central BU account. Your department’s usage will be tracked and billed internally. Budgeting alerts and notifications can be set based on project requirements.

Key Features

• Wide Range of AI Models: You get access to leading frontier models, including those from OpenAI and Anthropic, and can choose the appropriate model based on requirements and budget for your project.
• Scalable, Fully Managed Infrastructure: The LLM services are cloud-hosted, so you don’t have to set up or maintain any servers yourself.
• Easy Integration & Developer Tools: You can interact with these models using standard APIs and libraries, making it straightforward to incorporate into your applications, websites, or research workflows. In short, if you can call a REST API or use an SDK in languages like Python or Java, you can use these LLM APIs. IS&T will also provide documentation and guidance to help you get started with calling the APIs and handling keys.

Requirements

• This service is available to Boston University faculty, researchers, and sponsored departments/projects. (At this time, it is not intended for students or external users.)
• Departmental Financial Approval: A department financial administrator’s approval is required to request an LLM API key. This ensures your department is aware of and agrees to cover the cloud usage costs associated with the API. Typically, you should coordinate with your department’s financial administrator or budget manager before requesting access.
• Appropriate Use Compliance: Users must agree to use these APIs in accordance with all applicable usage policies, data guidelines, and ethical rules set by the university and the vendors.

Getting Started

1. Obtain Department Approval: First, discuss your intent to use the LLM API service with your department’s financial administrator. Ensure that you have their approval and backing, as they will need to provide an internal billing account or budget number. It’s also a good idea to clarify your anticipated usage (e.g. which model and how intensively you plan to use it) so everyone is aware of potential costs.
2. Submit the Request Form: Fill out the AI API Key Request Form to formally request access. On the form, you’ll need to provide information such as the intended use case for the API, and confirmation of financial approval (e.g. entering the approving administrator’s name or uploading an approval email). Once submitted, the request will go to IS&T for review.
3. Provisioning and Access: After your request is reviewed and approved by IS&T, the team will provision the necessary API key(s) or credentials for you. You will receive instructions on how to access the keys (for example, you might get an Azure OpenAI key, an AWS Bedrock access configuration, or both, depending on what you requested). IS&T will also provide any available documentation on using the keys, including endpoint URLs, example code, and best practices for managing your API usage. At this point, you can integrate the keys into your applications or research scripts and begin making calls to the LLM services.