# Arize AI MCP for AI Agents AI Agent Connect > Arize AI MCP lets you manage machine learning observability and model monitoring directly through your AI agent. Instead of jumping between tabs to check for data drift or inspect execution spans, you can ask your agent to pull telemetry, list active projects, and manage datasets. It turns your AI into a dedicated ML operations assistant that keeps your models running smoothly without the manual overhead of a dashboard. ## Overview - **Category:** friends-mcp - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_wCwq9adpwddtNPc4Bp17r5LJCv19cGsWZBiDUv4U/ai-agent-connect - **Tags:** ml-observability, model-monitoring, data-drift, ai-performance, telemetry, troubleshooting ## Description The Arize AI MCP lets you manage machine learning observability and model monitoring directly through your AI agent. Managing machine learning models at scale usually means jumping between dozens of tabs to check for data drift, inspect telemetry, and verify that your production models are actually behaving as expected. This Connector changes that by bringing the entire Arize AI suite into your primary workspace. You can now talk to your agent to see how a specific project is performing or pull high-fidelity execution spans without ever leaving your code editor or chat window. It's about getting the answers you need in plain English rather than digging through nested menus in an observability portal. By connecting through the Vinkius catalog, you give your AI the context it needs to act as an ML operations coordinator. You can check model metadata, track experiments, and organize your datasets through simple conversation. It cuts out the friction of manual logging and lets you focus on the actual science of building and refining your models. ## Tools ### create_dataset Create a new dataset for model evaluation or validation. This lets you organize your data for testing without manual portal entries. ### get_model Get specific details about an ML model. Use this to quickly check metadata and coordinate your AI strategy. ### list_datasets List all existing datasets in your account. This helps you keep track of your evaluation data at a glance. ### list_experiments List all ML experiments. This is useful for comparing results across different environments and runs. ### list_projects List all active tracing projects. Use this to see which projects are currently being monitored. ### list_spans List execution spans for a project. This provides high-fidelity telemetry data to find specific performance bottlenecks. ## Prompt Examples **Prompt:** ``` List all active ML projects in my Arize account. ``` **Response:** ``` I've retrieved your tracing projects. You currently have 3 active projects, including 'Production Classifier' (ID: 1024) and 'Beta Recommender'. Which one would you like to inspect for recent traces? ``` **Prompt:** ``` Show the recent execution spans for project '1024'. ``` **Response:** ``` Accessing telemetry data... I found 5 recent spans for project 1024. Most executions are showing low latency (avg 120ms). One span is flagged with a 'Schema Mismatch' warning. Shall I retrieve the detailed metadata for that trace? ``` **Prompt:** ``` Create a new dataset 'Q2_Eval_Data' for model evaluation. ``` **Response:** ``` Dataset orchestrated! I've successfully created 'Q2_Eval_Data' in your Arize environment. The dataset ID is 'ds_456'. You can now begin uploading model versions for automated high-fidelity validation. Shall I list your available models? ``` ## Capabilities ### Pull telemetry data Get high-fidelity execution spans for specific projects to see exactly where a model might be lagging. ### Manage ML datasets Create and organize datasets for evaluation and validation without manual entry. ### Track experiments View a list of experiments to compare model performance across different environments. ### Inspect model metadata Retrieve specific details about your models to coordinate your AI strategy. ### Monitor active projects List all your tracing projects to get a bird's-eye view of your production environment. ## Use Cases ### Identifying a production bottleneck An engineer asks the agent to list spans for a project and identifies a specific execution with a schema mismatch. ### Organizing evaluation data A data scientist wants to create a new dataset for Q2 data and uses the agent to create it instantly. ### Checking model metadata A developer needs to know the metadata for a specific model to see if it's ready for deployment. ### Auditing active tracing projects An ML lead wants to see all active tracing projects to ensure the team is following the correct monitoring plan. ## Benefits - Stop manual dashboard hopping by using `list_projects` and `list_spans` to get telemetry directly in your chat. - Manage your ML infrastructure faster by using `create_dataset` to build evaluation sets without leaving your workspace. - Get instant visibility into model health by calling `get_model` to pull metadata for specific production assets. - Track your research progress more easily by using `list_experiments` to compare different model versions. - Troubleshoot production errors quickly by pulling high-fidelity execution spans with `list_spans` to identify failures. ## How It Works The bottom line is you get a way to query your ML observability data using natural language instead of a dashboard. 1. Subscribe to the Arize AI MCP on Vinkius. 2. Grab your API Key from the Arize dashboard under Settings. 3. Connect the key to your AI client to start querying your ML telemetry. ## Frequently Asked Questions **Can Arize AI help me find data drift?** Yes, it allows your AI agent to query your model performance and identify drift points without you having to manually filter through complex dashboards. **How do I check my model performance with Arize AI?** You can ask your agent to retrieve specific model metadata or list recent execution spans to see how your models are performing in real time. **Can I manage my ML datasets using Arize AI?** Yes, you can use your AI agent to create and list datasets for your evaluation and validation workflows through simple natural language commands. **What kind of telemetry does Arize AI provide?** It provides high-fidelity execution spans and telemetry data, which helps you see exactly how your models are behaving in production. **Can Arize AI help me troubleshoot production errors?** Absolutely. Your agent can pull specific span details for any project, making it much faster to pinpoint and fix production bottlenecks. **Is Arize AI good for tracking ML experiments?** Yes, it lets you list and monitor all your active experiments, making it easy to compare results across different environments. **How do I find my Arize API Key?** Log in to your account, navigate to **Settings** > **API**, and generate or copy your unique secret key. **Can I track model drift via AI?** Yes! Use the `list_experiments` tool to retrieve data on active model evaluations and track performance variations programmatically. **How do I retrieve telemetry traces?** Use the `list_spans` tool to retrieve high-fidelity execution spans and traces for your ML projects directly from the platform.