Connecting Applications to LiteLLM Sprinted Updated v1
Connecting Applications to OU LiteLLM
Models, API Keys & Tool Connections
University of Oklahoma Libraries
Digital Scholarship & Data Services (DSDS)
Last Updated: May 2026
Table of Contents
- 1 Table of Contents
- 2 How OU LiteLLM Works
- 3 1. Model Names
- 3.1 AWS Bedrock Models
- 3.2 NRP Models
- 4 2. API Endpoints
- 5 OpenAI-Compatible API Configuration
- 6 3. Create API Keys
- 6.1 Step-by-Step
- 6.2 Test Your API Key
- 7 4. Connecting LiteLLM to Tools
- 7.1 AnythingLLM
- 7.2 Chatbox AI
- 7.3 Claude Code & Cowork
- 7.4 Liquid Apollo
- 7.5 OpenCode
- 7.6 Open WebUI
- 8 5. Using LibreChat with OU Models (No-Code Chatbot Platform)
- 9 6. Support & Resources
- 9.1 Contact
- 9.2 Quick Links
About This Guide This guide explains how to connect third-party applications and AI tools to the OU LiteLLM gateway. It includes supported model names, API key setup, and configuration examples for common desktop, mobile, and web-based AI applications. |
Who This Guide Is For
Researchers
Students
Developers
Users connecting AI tools without coding
How OU LiteLLM Works
OU LiteLLM provides a single OpenAI-compatible API endpoint that routes requests to multiple AI model providers.
TODO: Add LiteLLM User Guide link
1. Model Names
Use the exact API model name when making API calls or configuring tools. Display names are for your reference only — the API requires the string in the “API Model Name” column.
AWS Bedrock Models
These models are commercial and funded by OU Libraries. They are accessed through the Bedrock endpoint.
Display Name | API Model Name |
Amazon Nova Lite | amazon.nova-lite-v1:0 |
Claude 3 Haiku | anthropic.claude-3-haiku-20240307-v1:0 |
Claude Sonnet 4.5 | global.anthropic.claude-sonnet-4-5-20250929-v1:0 |
NRP Models
These models run on the National Research Platform. The API name (left column) is what you enter in your code. The full model name (right column) shows what’s running behind the scenes.
API Model Name | Full Model Name |
qwen3 | Qwen3.5-397B-A17B-FP8 |
gpt-oss | gpt-oss-120b |
kimi | Kimi-K2.6 |
gemma4 | gemma-4-31B-it |
gemma4-small | gemma-4-E4B-it |
glm-4.7 | GLM-4.7-FP8 |
minimax-m2 | MiniMax-M2.7 |
Qwen3 Coder 30B | Qwen3 Coder 30B (code-optimized) |
Qwen3 32B | Qwen3 32B (reasoning-optimized) |
Check NRP model availability: nrp.ai/llm-status
2. API Endpoints
LiteLLM uses one base URLs depending on the provider. Always use the correct endpoint for the model you’re calling.
OpenAI-Compatible API Configuration
Most third-party applications connect successfully using OpenAI-compatible or Generic OpenAI settings.
Setting | Value |
Base URL | |
Authentication | Bearer Token |
API Key | Your LiteLLM API key |
Model Name | Exact API model name |