- HEALTH
- PROGRESS
- STATUS
- HUMAN
Short answer
How can I monitor my AI model's performance?
Your AI client queries Arize AI to pull project lists, specific experiment metrics, and model details. This allows you to spot data drift or performance drops without leaving your chat interface.
The outcomes
What you can track
The AI pulls specific data points from your Arize AI workspace.
HEALTH
Identify model drift
The AI checks for shifts in your data distribution. You'll see exactly when your model starts behaving differently than it did during training.
PROGRESS
Track experiment runs
Your agent pulls metrics from active experiments. You can see how different versions are performing side-by-side.
STATUS
Check project health
Get a high-level view of your active projects. The AI identifies which models need your immediate attention.
HUMAN
Decide on retraining
The AI provides the data, but you decide when to trigger a new training run or adjust your deployment strategy.
The workflow
What your AI does when the data arrives.
Your agent acts as a bridge between your chat interface and your Arize AI telemetry.
Find your projects
The AI scans your workspace to find the relevant projects you want to monitor.
list_projectsInspect experiment metrics
Once a project is identified, the AI pulls the specific metrics from your latest runs.
list_experimentsAnalyze model details
The AI pulls specific configuration and performance data for a single model.
get_modelReport on drift
The AI compares current data against your baseline to flag performance issues.
list_experiments
Try it
Copy these to start.
Use these prompts to begin monitoring your models.
Starting points
Swap out the project or experiment names with your actual Arize AI identifiers.
Arize AI Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_wCwq9adpwddtNPc4Bp17r5LJCv19cGsWZBiDUv4U/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 — Arize AI capabilities are ready to use.
{
"mcpServers": {
"arize-ai-alternative-mcp": {
"url": "https://edge.vinkius.com/vk_preview_wCwq9adpwddtNPc4Bp17r5LJCv19cGsWZBiDUv4U/mcp"
}
}
}List all my active projects in Arize AI.
Show me the metrics for the 'customer-churn-v2' experiment.
What is the current status of the 'recommendation-engine' model?
Check the latest experiment results for my fraud detection project.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Arize AI once, then ask.
One connection gives your agent full access to your Arize AI telemetry. Your credentials stay encrypted and secure.
Connect Arize AI to your AIFAQ
How the monitoring works
- 01
Can the AI fix model drift automatically?
No. The AI can only report on drift and provide the data you need to make a decision.
- 02
Can I delete projects through the AI?
No. The tools only allow for reading project, experiment, and model data.
- 03
Does the AI see all my data?
The AI only accesses the specific project and experiment data you ask it to query.
- 04
Can the AI trigger a new training job?
No. The AI is limited to retrieving existing metrics and model information.
- 05
How often does the AI check for drift?
The AI only checks for drift when you specifically ask it to query your experiment or model data.
More questions about Arize AI? The Connector page answers them. See everything the Arize AI Connector can do
Connect Arize AI to Claude, Cursor, ChatGPT & more
