AI Tools and Integrations
Research Computing supports various AI tools and hardware to accelerate your research.
One guide takes you from nothing to a working setup: an API key, an AI coding assistant that can read and edit your projects, and a working sense of how to get good results from the models we host. No prior experience assumed.
Research Computing Offerings
Hardware
We provide high-performance hardware for AI workloads:
| Platform | Description |
|---|---|
| Gaudi Accelerators | Intel Gaudi2 with 8x interconnected HPUs per node. ~200 nodes available (10 on Slurm, remainder on Kubernetes). |
| NVIDIA GPUs | Industry-standard GPUs for AI/ML workloads including A100 and H100. |
| AMD GPUs | High-performance AMD accelerators for compute workloads. |
Portals
Access AI tools through the following web portals:
| Portal | URL | Description |
|---|---|---|
| OpenWebUI | ai.rc.asu.edu | A user-friendly interface for interacting with LLMs, similar to ChatGPT. |
| Sim Studio | agentic.rc.asu.edu | An open-source platform to build and deploy AI agent workflows visually. |
| Jupyter AI | sol.asu.edu | Integrated AI coding assistant within the JupyterLab environment. |
API Access
Access Large Language Models programmatically for your applications and scripts. See the LLM API Access guide for details on requesting API keys and usage examples.
AI coding assistants
Assistants that read your project and edit files for you, all connected to the RC gateway with your own key:
| Guide | What it covers |
|---|---|
| Getting Started with AI | The complete setup path, plus agents, skills, and prompting. Start here. |
| Set Up OpenCode | The recommended assistant — terminal and desktop app. |
| Connect VS Code to the RC API | VS Code Chat with the RC gateway (BYOK). |
Skills and agent configs
The Voyager portal LLM Access tab includes an opencode config generator (presets, agents, and permissions, no hand-edited JSON) and a Skills library of reusable HPC and research instructions your agents load on demand — Slurm scripts, job troubleshooting, GPU right-sizing, Apptainer containers, reproducible research, and more. See Getting Started with AI.
Getting Help
If you encounter issues or need assistance with the AI tools, please refer to our Getting Help page to submit a support request or view office hours.