LLM Gateway

The URI LLM Gateway is a centrally managed service from URI Information Technology Services that provides the URI community with API access to many large language models (LLMs) through a single endpoint. Instead of setting up separate accounts with individual AI vendors, one gateway key works with models from Anthropic, Meta, Mistral, Qwen, NVIDIA, and others.

The gateway is designed for people who want to work with AI beyond a chat window — using AI coding agents and command-line tools such as OpenCode, connecting code editors like VS Code, writing scripts, or building applications such as chatbots, document analysis tools, and automated workflows. It also powers other URI services, including LibreChat and the Enrollment Services and College of Pharmacy AI assistants. Looking for a ready-to-use chat interface instead? See LibreChat.

Contact

ITS Cloud and IT Infrastructure
its-citi@uri.edu

Features

  • One key, many models: Access 20+ large language models from multiple vendors with a single API key.
  • OpenAI-compatible: Works with any tool or library that supports the OpenAI API, including Python libraries, code editor extensions, and command-line tools.
  • Range of models: Choose from fast, lightweight models for simple tasks up to the most capable models for complex reasoning and analysis.
  • Protected data: Models run within URI’s cloud environment, and prompts and responses are not used to train AI models.
  • No vendor lock-in: Switch models by changing a single setting — no new accounts or software needed.
  • Centrally managed: Access and data policies are handled by ITS.

Access

The LLM Gateway is available to the entire URI community. To request an API key, email its-citi@uri.edu. Keys are issued to individuals and should not be shared.

Keep your key secure: do not commit it to code repositories or paste it into shared documents or chat messages. Store it in an environment variable or a protected secrets file.

Support and Training

ITS maintains the LLM Gateway and can help with access, API keys, and connecting your tools — including Python, code editors such as VS Code, and other OpenAI-compatible clients. For assistance, email its-citi@uri.edu.

Tip: Start with a lighter, faster model and move to a more capable model only if you need better results.