Premium Feature — AI-backed features are available in SM Professional, LIMS Starter, LIMS Enterprise, Biologics LIMS, SDMS Professional, and SDMS Enterprise.
Learn more or
contact LabKey.
AI features depend on several components working together behind the scenes: an AI model, the services that feed it relevant information, and the safeguards around it. The
AI Assistant Status page gives administrators a single dashboard for these components, so they can confirm that AI features are set up and protected. This topic explains how to open the page and what each row shows.
View the AI Assistant Status Page
- Select > Site > Admin Console.
- Under Diagnostics, click AI Assistant Status.
Status Rows
The page shows one row for each component:
- MCP server: Whether the MCP server is enabled.
- AI guardrails: Whether a guardrail is protecting AI Assistant requests. Shows "enabled" or "disabled". See AI Guardrails below.
- LLM model: The large language model that answers AI Assistant requests.
- Embedding: The model that converts text into a searchable form, so the AI Assistant can find relevant content.
- Vector Store: Where that searchable content is stored.
- MCP sessions: The sessions currently connected to the MCP server.
AI Guardrails
A guardrail is a safety filter that screens AI Assistant requests, much like a spam filter screens incoming email. The
AI guardrails row confirms that a guardrail is in effect. What "enabled" means depends on where your server is deployed:
| Status | LabKey Cloud | On-Premise |
|---|
| Enabled | A guardrail is configured and was confirmed working at server startup. | A guardrail ID is configured. LabKey doesn't verify that it works. |
| Disabled | No guardrail is configured, or it couldn't be confirmed. In that case, AI features are turned off. | No guardrail ID is configured. |
Related Topics
