Clemson RCD LLM

Language model provider for the Clemson RCD LLM Service. This extension makes self-hosted language models from Clemson University's Research Computing division available directly in VS Code's chat. See the documentation for more information.

Quick Start

Install this extension
Get an API key from the RCD LLM dashboard
Open the Command Palette (Ctrl+Shift+P or Cmd+Shift+P) and run RCD LLM: Set API Key
Enter your API key when prompted
Open the chat panel and select a model from the "Clemson RCD" provider

Sandboxing

We encourage you to use agent sandboxing. This uses OS primitives to block actions on your computer without needing approvals and human inspection. It can filter filesystem access and network access.

This setting defaults to off. To enable it and see all the other sandbox settings you can use, see the VS Code documentation on sandbox agent commands.

Configuration

Open VS Code Settings (Ctrl+, or Cmd+,) and search for "RCD LLM".

Setting	Default	Description
`rcdLLM.showExternalModels`	`false`	Show externally hosted models (e.g. OpenAI) in the model picker.
`rcdLLM.showOfflineModels`	`false`	Show models currently offline.

Status Bar

A status bar item in the lower-left shows the current connection state.

Icon	Text	Tooltip	Meaning
`$(sync~spin)`	RCD LLMs	Checking connection...	Checking connection to the service
`$(cloud)`	RCD LLMs	Connected • N models available	Connected, N models available
`$(warning)`	RCD LLMs	No API key configured. Click to set up.	No API key configured
`$(error)`	RCD LLMs	Error:	Connection failed (e.g. invalid key, service down)

Click the status bar item to see status details and access quick actions:

Set API Key — store or update your API key
Clear API Key — remove the stored key
Configure Settings — open RCD LLM settings
View Documentation — open the RCD LLM docs in your browser
Refresh Models — force-refresh the model list

The model list refreshes automatically every hour. To refresh manually, run RCD LLM: Refresh Models from the Command Palette (Ctrl+Shift+P or Cmd+Shift+P).

Commands

Open the Command Palette (Ctrl+Shift+P or Cmd+Shift+P) and search for "RCD LLM":

Command	Description
RCD LLM: Set API Key	Store your API key securely in VS Code's SecretStorage
RCD LLM: Clear API Key	Remove the stored API key
RCD LLM: Refresh Models	Force-refresh the model list from the service
RCD LLM: Show Status	Show connection status and quick actions

Model Picker

Models are sorted with alias models first. Aliases are friendly names that point to a specific hosted model and are kept up to date by the RCD team. For example, tiger always points to the most powerful model currently available.

Each model shows a detail line with useful metadata:

Alias for X — this is an alias pointing to model X
experimental / deprecated — lifecycle status (non-active models only)
retires M/D/YYYY — retirement date for deprecated models
RCD AI Cluster / OpenAI — which provider serves the model

Hover over a model to see its full description and license information.

Troubleshooting

No models appear

Check the Output panel (View > Output) and select RCD LLM from the dropdown
Ensure your API key is set: Command Palette (Ctrl+Shift+P or Cmd+Shift+P) → RCD LLM: Set API Key
Run RCD LLM: Refresh Models from the Command Palette to retry

Support

If you are still having problems, please see our support page for ways to contact the RCD team.

Logs

View extension logs in the Output panel (View > Output) by selecting RCD LLM from the language dropdown. Logs include model discovery, filtering, and chat request details.