- STATUS
- METRICS
- DETAILS
- HUMAN
Short answer
How can I check my Predibase model status using AI?
Your AI pulls endpoint health and performance data directly from Predibase into your chat. You'll get instant updates on model status and specific usage metrics. This breakdown shows how it works.
Where your data lands.
Deployment insights.
The AI pulls information from Predibase and presents it in your chat.
STATUS
Live endpoint checks
The AI checks if your models are running or down. You get immediate clarity on deployment availability.
METRICS
Performance snapshots
Your agent pulls usage and performance numbers. You can see how your models behave without opening a dashboard.
DETAILS
Model configuration
The AI retrieves specific information about your deployed models. This helps you verify settings on the fly.
HUMAN
Manual intervention
The AI identifies issues but cannot change your deployment settings. You stay in control of all model updates.
The workflow
What your AI does when the request arrives.
The AI acts as a bridge between your chat interface and your Predibase deployments.
Check availability
The AI looks at your active endpoints to see if they are healthy and responding.
get_healthGather performance data
Your agent pulls real-time metrics to show how your models are performing.
get_metricsRetrieve model details
The AI pulls specific information about your model configurations and deployment settings.
get_infoReport findings
The AI summarizes the health, metrics, and info into a plain text response for you.
get_healthget_metricsget_info
Try it
Copy these to start.
Use these prompts to begin monitoring your deployments.
Starting points
These are opening lines. Swap out the model names or endpoint IDs with your actual Predibase identifiers.
Predibase Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_anUwoOMv2E6QQVudvBXsDytO30kiB1fpaXuzKUbp/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Predibase capabilities are ready to use.
{
"mcpServers": {
"predibase-llm-serving-finetuning-mcp": {
"url": "https://edge.vinkius.com/vk_preview_anUwoOMv2E6QQVudvBXsDytO30kiB1fpaXuzKUbp/mcp"
}
}
}Check the health of my production endpoint.
What are the current performance metrics for my fine-tuned model?
Get the deployment info for my latest LLM endpoint.
Are any of my Predibase endpoints currently reporting issues?
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Predibase once, then ask.
You only need to link your account once. Your credentials stay encrypted, and you can immediately start asking your AI about your models.
Connect Predibase to your AIFAQ
How this task behaves.
- 01
Can the AI fix a broken endpoint?
No. The AI can only check the health and report status. You have to manually fix any deployment issues in Predibase.
- 02
Can the AI change my model settings?
No. The tools only allow the AI to read health, metrics, and info. It cannot modify your fine-tuned models or deployment configurations.
- 03
Does the AI see my model weights?
No. The AI only accesses the metadata and performance metrics provided by the tools.
- 04
How often does the AI check metrics?
The AI only checks metrics when you specifically ask it to via a prompt.
- 05
Can I use this with any AI client?
Yes, as long as your client is MCP-compatible, like Claude, Cursor, or Windsurf.
More questions about Predibase? The Connector page answers them. See everything the Predibase Connector can do
Connect Predibase to Claude, Cursor, ChatGPT & more
