Starburst MCP. Query federated data lakes from natural conversation.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Starburst MCP Server connects your AI client to Starburst Enterprise. It lets you run complex SQL queries across massive, federated data lakes—meaning you can query sources like Hive, Snowflake, and Postgres without ever switching database clients.
You use this to discover schemas, manage roles, and execute cross-platform analytics from a single prompt.
What your AI agents can do
Get query details
Retrieves detailed information for a single, specified SQL query.
List catalogs
Lists all connected data catalogs available in the Starburst Galaxy instance.
List data products
Finds and lists all published, curated datasets across the entire Starburst network.
The AI client executes complex, federated SQL against multiple data sources (like Hive or Snowflake) in a single step.
You can list every catalog available in the Starburst Galaxy using list_catalogs to see what systems are hooked up.
The tool allows you to browse published data products (list_data_products) and group them into logical domains (list_domains).
You can list all existing security roles using list_roles to confirm who has access to what.
Use list_queries to see a record of the most recent SQL queries run in the cluster.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
Starburst MCP Server: 6 Tools for Data Discovery & Querying
Use these tools to list connected catalogs, manage data products, check security roles, and execute complex SQL queries against your entire federated data lake.
019d760dget query details
Retrieves detailed information for a single, specified SQL query.
019d760dlist catalogs
Lists all connected data catalogs available in the Starburst Galaxy instance.
019d760dlist data products
Finds and lists all published, curated datasets across the entire Starburst network.
019d760dlist domains
Groups and lists data product domains, which helps organize large sets of related data assets.
019d760dlist queries
Shows a list of the most recent SQL queries that have been run against the cluster.
019d760dlist roles
Lists all active security roles configured within the organization's access control system.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Starburst, then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
What you can do with this MCP connector
You're connecting your AI client straight into Starburst Enterprise, which means you can run complex SQL queries across massive, federated data lakes without ever jumping between different database clients like Hive or Snowflake. You use this setup to execute cross-platform analytics and discover schemas from a single prompt.
The system lets your agent perform the heavy lifting: it executes advanced, federated SQL against multiple sources in one step, giving you structured analysis right where you are working.
When you need to figure out what data exists, you start by mapping the environment. You use list_catalogs to see every single data catalog connected within your Starburst Galaxy instance, which tells you exactly which systems are hooked up for querying. Once you know the sources, you can get a sense of the overall organization by running list_domains, grouping together large sets of related data assets into logical domains.
To find specific datasets, you browse published data products using list_data_products. This tool finds and lists all curated analytical datasets across your entire network. You'll want to check this list before writing a query because it validates the definitions and boundaries of what's available for querying.
For those who need to audit access or confirm permissions, you can use list_roles to see every active security role configured in the organization's access control system. This lets you verify exactly who has clearance to run certain types of queries against sensitive data.
When it comes to tracking down specific data points or understanding what a query is actually doing, you have multiple options. You can use list_queries to pull up a record showing the most recent SQL queries that were executed on the cluster; this gives you an immediate history check of who's been running what.
If one of those past queries was complex or needs deeper review, you don't have to guess. You can specify the exact query and run get_query_details to retrieve detailed information about that single SQL statement, including its parameters and execution context. All these capabilities—from listing roles to getting specific query details—make sure your agent has everything it needs right at your fingertips.
How Starburst MCP Works
- 1 Install the Starburst MCP connector and provide your
STARBURST_HOSTandSTARBURST_TOKENin the MCP settings to establish a connection. - 2 Tell your AI client what you need. For example: "List all data catalogs, then find any schemas related to finance."
- 3 The agent runs the necessary tools (like
list_catalogsandlist_schemas) and returns the structured results directly in the chat window.
The bottom line is that you skip the manual steps of connecting, validating credentials, and navigating multiple database UIs; your AI client does it all for you.
Who Is Starburst MCP For?
This tool is built for data people who spend too much time figuring out where the data lives before they can actually analyze it. It's for the Data Analyst drowning in source names, the Data Engineer managing dozens of connection strings, or the Governance Manager who needs to prove exactly who has access to what.
Uses the AI client to construct and run complex SQL queries across different data sources (e.g., mixing sales data from Snowflake with customer records in Hive) without writing boilerplate connection code.
Audits the entire data environment by listing catalogs (list_catalogs) and schemas to map out dependencies, or checking query history using list_queries.
Verifies access control by running list_roles and checking specific data product permissions to ensure compliance across the enterprise.
What Changes When You Connect
- Run complex cross-source queries instantly. You don't have to manually build connection strings or switch between database clients; the AI handles it using
execute_queryand your prompt. - See your entire data landscape at a glance. Using
list_catalogsquickly shows you every connected system (like Snowflake, Hive, etc.) without deep platform knowledge. - Manage access risk easily. Running
list_roleslets you confirm who has what permissions before executing anything sensitive—a huge win for governance. - Discover curated data assets fast. Instead of guessing which tables to check, use
list_data_productsto find vetted, published datasets immediately. - Track activity without opening a dashboard. You can review recent work by calling
list_queries, giving you an instant audit trail right in your chat.
Real-World Use Cases
The Cross-Source Audit
A Data Engineer needs to check if the 'production' customer IDs from Snowflake match up with the corresponding records in the legacy Hive warehouse. Instead of writing two separate scripts and running them manually, they prompt the agent: "Run a join query comparing top 10 customers between Snowflake and Hive.". The agent uses execute_query against both sources simultaneously, delivering one consolidated result.
The Missing Schema Check
A Data Analyst hears about a new 'finance' dataset but can't find it. They start by calling list_catalogs, which shows them the main source databases. Then, they ask the agent to inspect all schemas within the finance catalog using list_schemas (via an implicit tool call), immediately pinpointing the correct table structure.
The Compliance Review
A Governance Manager needs to know if a new team member has proper access. They run list_roles and then verify that the 'analyst' role only grants read-only access to specific data domains (list_domains), confirming compliance without manual checks.
The Debugging Query
A Data Scientist runs a complicated query that fails. Instead of wading through logs, they use list_queries to see the exact parameters of the last run, and then use get_query_details to check the original definition, solving the bug in minutes.
The Tradeoffs
Assuming a single source
Running a complex query that only targets one database (e.g., 'SELECT * FROM table') when the data actually needs to be pulled from three different, connected sources.
→ Don't write the SQL yourself; prompt your agent: "Execute a federated query joining customer records from Snowflake with transaction logs in Hive." The agent handles the multi-source structure.
Running queries without checking permissions
Writing and submitting a high-privilege SELECT statement immediately, only to receive an 'Access Denied' error halfway through.
→
Always run list_roles first. Then, confirm the role you need before executing any query.
Forgetting which data is published
Pointing your AI client at a raw operational table that isn't meant for general use, leading to messy and incorrect results.
→
Use list_data_products first. This tool only shows you datasets the organization has formally marked as ready for consumption.
When It Fits, When It Doesn't
You should use this server if your data needs live querying across multiple, disparate systems—think combining financial data from an old mainframe with modern web logs in a cloud warehouse. It’s necessary when you need to treat all your connected databases as one giant pool of information.
Don't use it if your task is simple: If you only need to run a query against one specific, local database instance and there are no cross-system dependencies, you might be better off using that database’s native client. Also, if your goal is just documentation review (like reading an old PDF report), this tool won't help—it deals with live, structured data only.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Starburst. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This server provides 6 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Sifting through a modern data lake feels like navigating a maze of acronyms and connection strings.
Today, finding the right data is a chore. You open the data portal, see dozens of catalogs—Hive here, Postgres there—and you have to manually check three different interfaces just to figure out which source holds the 'customer' record and if it’s even clean enough to use. It’s copy-pasting connection details and guessing where to look.
With this MCP server, all that manual effort disappears. You tell your AI client: "Show me all active catalogs and then list schemas containing anything related to finance." The agent handles the cross-system plumbing for you, giving you a clean, actionable map of every asset.
Starburst MCP Server: Run complex queries across diverse data sources.
Before this, getting a full picture meant writing multiple scripts and manually joining the results in Excel or a BI tool. It was slow, error-prone, and required deep knowledge of every single source's unique syntax.
Now, you just ask your AI client to join them. The agent uses `execute_query` and manages the federated query logic across all sources automatically. You get the final answer—not a list of connection errors.
Common Questions About Starburst MCP
How do I check what catalogs Starburst can see using list_catalogs? +
You call list_catalogs. This tool lists every connected data source (like Snowflake or Postgres) that the Starburst Galaxy is currently linked to. It's your master inventory.
Can I use list_data_products to find a clean dataset? +
Yes, list_data_products lets you see only the datasets that have been formally published and vetted by data stewards. This is much safer than querying raw tables.
What if I want to know who can run a query? Do I use list_roles? +
You use list_roles. This tool lists all security roles, letting you verify the access limits for any user or team before writing code.
Does get_query_details help if my query fails? +
Yes. If a query runs but gives vague errors, get_query_details lets you pull up the full, precise details of that specific SQL execution attempt for debugging.
I ran a complex query; how do I check its performance details using `get_query_details`? +
It fetches detailed metadata about a specific SQL execution. You can review the query's resource usage, optimization plan, and run duration to pinpoint bottlenecks for better future queries.
What is the best way to see all the supported data sources available using `list_catalogs`? +
This tool lists every connected data catalog in your federated environment. It confirms compatibility with various source types, including Snowflake, S3, and PostgreSQL.
If I need to narrow down my search within a dataset, how do I use `list_domains`? +
It lists the distinct domains attached to any data product. You use this to scope your query immediately, ensuring you only analyze relevant business areas.
How can I review my team's recent activity and check executed SQL queries using `list_queries`? +
This tool provides a history of recently run SQL statements on the cluster. It helps audit data access patterns, track usage volume, and verify who ran what query.
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).
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Shippo
Get the best shipping rates from top carriers, print labels, and track deliveries with an API that simplifies e-commerce shipping.
USAJOBS (OPM)
Search and analyze federal job opportunities directly from USAJOBS. Access active listings, historic data, and detailed job requirements.
Worksection
Collaborative project management to manage tasks, projects, and timers with AI using Admin API v2.
You might also like
Printful
Access Printful's product catalog — list products, inspect variants, and retrieve detailed size guides directly from any AI agent.
TrackingTime
Track how your team spends time across projects and clients with timesheets, reports, and budgeting tools that keep work profitable.
Zoho Billing Alternative
Automate your subscription management with Zoho Billing — manage organizations, products, and plans directly from your AI agent.