# Materialize MCP for AI Agents AI Agent Connect > Materialize MCP lets you manage streaming SQL databases directly through your AI agent. You can run SQL queries on live data feeds, spin up compute clusters, and monitor your pipeline health without switching tabs. It's built for anyone who needs to handle real-time data at scale without the usual friction of manual database management. ## Overview - **Category:** databases - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_b3YATqNUmev2sLGakCv9DXqsXBiI2C8N6xSKsiWQ/ai-agent-connect - **Tags:** streaming-sql, real-time-data, materialized-views, data-engineering, cloud-database ## Description Managing real-time data pipelines usually feels like a constant game of whack-a-mole. You're jumping between terminal windows to check if a cluster is healthy, then switching over to a SQL editor to run a query, and finally heading to a dashboard to see if the data actually arrived. This Connector changes that by bringing your streaming database management directly into your AI agent. Instead of hunting for the right command or dashboard, you just tell your agent what you need. You can spin up new compute resources to handle a spike in traffic or run complex queries against live feeds to see your data as it happens. It handles the heavy lifting of interacting with your streaming infrastructure, so you can focus on the actual data. You can debug a slow materialized view or check the current status of your environment through a single conversation. It's one of the practical tools in the Vinkius catalog that actually saves you from the tab fatigue of modern data engineering. You get a direct line to your streaming SQL environment, making it much easier to maintain high-velocity data pipelines without the constant context switching. It removes the friction of manual configuration and gives you a way to interact with your live data feeds using plain language. You can check your instance health, manage your compute clusters, and execute queries without ever leaving your primary workspace. It's about getting your hands on your data faster and with fewer clicks. ## Tools ### execute_sql Run standard SQL or Materialize-specific commands on your live data feeds. This lets you interact with your streaming data without opening a separate editor. ### check_health Instantly check if your Materialize instance is healthy and running. This helps you monitor your pipeline status and spot issues before they cause downtime. ### list_clusters Get a full list of all your available compute clusters in one go. This makes it easy to audit your environment and see what's currently running. ### create_cluster Create a new compute cluster with specific sizes like xs, s, m, l, or xl. This lets you scale your processing power instantly to handle data spikes. ## Prompt Examples **Prompt:** ``` Check if my Materialize instance is up and running. ``` **Response:** ``` I've checked the status for you. Your Materialize instance is currently healthy and operational. Everything looks good. ``` **Prompt:** ``` What clusters do I have right now? ``` **Response:** ``` You currently have 2 compute clusters active: - **analytics-cluster** (size: m) - **default** (size: xs) ``` **Prompt:** ``` Create a new source from my Kafka topic 'orders'. ``` **Response:** ``` I've executed that for you. The SQL statement to create the 'orders' source was successful, and Materialize is now ingesting data from that Kafka topic. ``` ## Capabilities ### Create new compute clusters Spin up new clusters with specific sizes like xs, s, m, l, or xl to handle data spikes. ### Run SQL on live data feeds Execute standard SQL and streaming commands directly against your live data. ### Check your instance health Get an instant status update on your database health and operational status. ### List active clusters See a full list of your available compute clusters and their configurations. ## Use Cases ### Scaling for a traffic spike An engineer sees a lag in data ingestion and asks the agent to use `create_cluster` to spin up an 'xl' instance. This allows them to handle the load immediately without manual intervention. ### Debugging a broken source A data engineer asks the agent to `execute_sql` to `CREATE SOURCE` from a new Kafka topic to see if the connection works. This helps them isolate the problem quickly. ### Morning health check A DevOps lead asks the agent to `check_health` every morning to ensure all pipelines are operational. It saves them from clicking through multiple dashboards to verify the status. ### Inventory audit An analytics lead wants to see what's running and asks the agent to `list_clusters` to audit current resource usage. This helps them stay on top of their environment's costs and configuration. ## Benefits - You can run queries on live data feeds without leaving your chat window by using `execute_sql`. This lets you see your data as it happens without having to open a separate SQL editor or browser tab. - Spin up new compute clusters instantly with `create_cluster` to handle unexpected traffic spikes. This replaces the need for manual configuration when you need to scale your processing power on the fly. - Get immediate status updates on your instance health using `check_health`. This helps you spot issues before they break your pipelines by giving you a quick way to verify everything is running smoothly. - Audit your active resources quickly by using `list_clusters` to see all current configurations. This gives you a clear view of your environment so you can manage multiple workloads without getting lost in the details. - Simplify complex streaming SQL commands by letting your agent handle the `execute_sql` syntax. This reduces the risk of typos and helps you get your queries right the first time when working with live feeds. ## How It Works The bottom line is you get a natural language interface for your entire streaming SQL infrastructure. 1. Connect your Materialize API Key to the Connector. 2. Open your AI client like Claude or Cursor. 3. Ask your agent to run SQL, check health, or manage clusters. ## Frequently Asked Questions **What can the Materialize MCP do for my data engineering workflow?** It lets you manage your streaming database using natural language. You can run SQL, manage clusters, and check health without leaving your agent's chat window. **Can I use the Materialize MCP to run SQL on live Kafka streams?** Yes, you can use it to execute SQL commands like creating sources from Kafka topics, allowing you to interact with your live data feeds immediately. **How does the Materialize MCP help with scaling my compute resources?** It allows you to spin up new compute clusters with specific sizes like 'm' or 'xl' on the fly, making it easy to handle traffic spikes without manual setup. **Is the Materialize MCP good for monitoring my database health?** It's perfect for quick health checks. You can ask your agent to verify the status of your instance at any time to ensure your pipelines are running smoothly. **Can I manage multiple Materialize clusters with this Connector?** Yes, you can use it to list all your available compute clusters, giving you a clear overview of your entire streaming SQL environment in one place. **How do I connect the Materialize MCP to my AI agent?** You just need to subscribe to the Connector and provide your Materialize API Key. From there, your agent can handle all your streaming SQL tasks. **Can I create a new materialized view using this server?** Yes. You can use the `execute_sql` tool to run any valid Materialize SQL command, including `CREATE MATERIALIZED VIEW` to start processing your data streams in real-time. **How do I scale my compute resources through the AI?** You can use the `create_cluster` tool and specify a size (xs, s, m, l, or xl). This allows you to provision new compute capacity directly through the conversation. **Is there a way to check if my Materialize instance is currently reachable?** Yes, the `check_health` tool is designed specifically for this. It returns the current status of your instance to confirm it is operational.