# Starburst MCP for AI Agents AI Agent Connect

> Starburst connects your AI to Starburst Enterprise. It lets your agent query federated data lakes, manage access roles, and explore complex Trino-based data environments using natural language. Stop switching between database clients and start asking your AI to find, summarize, and audit your data products directly.

## Overview
- **Category:** industry-titans
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_O3fkMMqaRDxXU8YcUOdnSQyixPMchBfro4bAKlB1/ai-agent-connect
- **Tags:** data-lake, federated-query, trino, data-engineering, sql-analytics, data-governance

## Description

You can finally stop jumping between different database clients to find what you need. Instead of writing long SQL strings or hunting through documentation, you can just ask your agent to find specific tables or summarize a dataset. This Connector connects your AI to Starburst Enterprise, giving it the ability to navigate federated data lakes and manage organizational roles. If you use Vinkius to manage your tools, you can get this connected in minutes. It handles the heavy lifting of talking to Trino-based data products so you can focus on the actual analysis. You get to see schemas, list data products, and check permissions without ever leaving your chat window.

## Tools

### get_query_details
Retrieves specific details for a single SQL query. It helps you see exactly what happened during a previous execution.

### list_catalogs
Lists all data catalogs available in Starburst Galaxy. Use this to see which sources like S3 or Snowflake are connected.

### list_data_products
Lists all published data products in the network. This makes it easy to see what data is ready for consumption.

### list_domains
Lists the different data product domains. Use this to understand how your data is organized across the enterprise.

### list_queries
Shows a list of recent SQL queries executed in the cluster. This helps you track activity and debug recent requests.

### list_roles
Lists all security roles in your organization. Use this to audit permissions and check access levels.

## Prompt Examples

**Prompt:** 
```
Can you show me all the catalogs we have connected and tell me which ones have a 'finance' designation?
```

**Response:** 
```
I've checked your connected catalogs. Here is what I found:

| Catalog Name | Source Type |
| :--- | :--- |
| `finance_db` | Snowflake |
| `ops_hive` | S3 |
| `marketing_redshift` | Redshift |

**Note**: The `finance_db` catalog is your primary source for all financial reporting data.
```

**Prompt:** 
```
List the published data products and tell me the schema for the 'customer_insights' one.
```

**Response:** 
```
Here are the currently published data products:

*   `Marketing Insights Dashboard`
*   `Global Sales Summaries`
*   `Customer Insights`

**Schema for `Customer Insights`**:
*   `customer_id` (UUID)
*   `signup_date` (Timestamp)
*   `lifetime_value` (Float)
*   `last_purchase_category` (String)
```

**Prompt:** 
```
Who has access to the production_hive catalog? Show me the roles.
```

**Response:** 
```
I've retrieved the security roles associated with that environment:

*   `data_governance_admin` (Full Access)
*   `analyst_tier_1` (Read Only)
*   `finance_auditor` (Restricted Read)

Would you like me to check the specific permissions for any of these roles?
```

## Capabilities

### Run SQL on federated data
Your agent executes complex queries across multiple data sources and returns the results in a readable format.

### Explore catalogs and schemas
You can ask the AI to map out your data landscape to find specific tables or databases.

### Manage data products
The AI can list and retrieve published data products across your entire Starburst network.

### Audit security roles
Check who has access to what by asking the AI to list and inspect organizational roles.

### Track query history
See what has been run recently by pulling a list of recent SQL queries from the cluster.

## Use Cases

### Finding specific finance data
An analyst asks the agent to find all tables related to 'finance' in the production catalog and summarize the latest transaction counts.

### Auditing security permissions
A governance manager asks the AI to list all active roles and identify which ones have access to the payroll schemas.

### Exploring new data products
A data engineer wants to see what new products were published last week and get the schema details for a specific marketing dashboard.

### Debugging recent queries
A developer asks the AI to show the last 10 queries run in the cluster to see if a specific job failed or succeeded.

## Benefits

- Skip the SQL boilerplate by using `execute_query` to run complex statements directly from your chat.
- Find tables faster by using `list_catalogs` and `list_schemas` to map out your data landscape instantly.
- Keep your data organized by using `list_data_products` to see what is published across the network.
- Stay compliant by using `list_roles` to check who has access to what in your organization.
- Track activity easily by using `list_queries` to see what has been run recently in the cluster.
- Get summarized results from complex queries without having to manually parse through hundreds of rows of data.

## How It Works

The bottom line is you get to talk to your data instead of writing SQL to find it.

1. Add your STARBURST_HOST and STARBURST_TOKEN to your Connector configuration.
2. Ask your agent to perform a task like finding a schema or running a specific SQL query.
3. Receive the structured data or a summarized report directly in your chat.

## Frequently Asked Questions

**Does the Starburst MCP work with my S3 data lake?**
Yes. It connects to your Starburst Enterprise environment, which allows it to query federated sources including S3, Snowflake, and Iceberg.

**Can I use Starburst MCP to manage my data roles?**
Yes. You can ask your agent to list and inspect organizational roles to help with auditing and access control.

**How does Starburst MCP help with data discovery?**
It lets you ask natural language questions to find specific catalogs, schemas, and published data products across your entire network.

**Is Starburst MCP compatible with Trino?**
Yes. It is designed to work with Trino-based data products to help you query and manage complex data environments.

**Can the Starburst MCP run SQL queries for me?**
Yes. Your agent can execute SQL statements directly and return structured results or summaries to you in the chat.

**Will Starburst MCP show me my recent query history?**
Yes. It can pull a list of recent SQL queries executed in the cluster, which is helpful for debugging or checking recent activity.

**Can the AI run write or drop queries?**
That depends on your token's permissions. The `execute_query` tool passes any SQL to Starburst. To prevent mutations, provision a read-only token via your Starburst security settings.

**How do I handle queries that return thousands of rows?**
Always add a `LIMIT` clause in your SQL. Large result sets can exceed the AI's context window. Instruct the agent to use `LIMIT 10` or `LIMIT 50` to keep responses manageable.

**What credentials do I need?**
Two values: `STARBURST_HOST` (your cluster URL, e.g., `https://mycluster.trino.galaxy.starburst.io`) and `STARBURST_TOKEN` (a service account token from your Starburst admin settings).