Spark SQL

aidancorrell/spark-sql-mcp-server

Connects Claude to any HiveServer2-compatible system (Spark, EMR, Hive, Impala) over the Thrift protocol. Exposes four tools: list_databases, list_tables, describe_table, and execute_query for read-only SQL operations. Enforces safety by blocking non-SELECT statements and automatically limiting unbounded queries. Supports multiple auth methods including LDAP, Kerberos, and NOSASL. Built for AWS EMR workflows with SSH tunnel support, though it works with any Spark cluster exposing port 10000. Useful when you need Claude to explore schemas and run analytics queries against production data lakes without write access. Credentials stay local via environment variables. Ships with Docker Compose setup and sample data for local testing.

tools: 4
auth: Secret needed
transport: STDIO
env vars: 7
last updated: Feb 8, 2026

View on GitHub

Tools

4 tools

Public tool metadata for what this MCP can expose to an agent.

list_databasesList all available databases

List all available databases

No parameter schema in public metadata yet.

list_tablesList tables in a database

List tables in a database

No parameter schema in public metadata yet.

describe_tableGet table schema (columns, types)

Get table schema (columns, types)

No parameter schema in public metadata yet.

execute_queryRun read-only SQL queries with formatted results

Run read-only SQL queries with formatted results

No parameter schema in public metadata yet.

Configuration

SPARK_HOST*

Hostname of the Spark Thrift Server

SPARK_PORT

Port of the Spark Thrift Server (default: 10000)

SPARK_DATABASE

Default database to use

SPARK_AUTH

Authentication method: NONE, LDAP, KERBEROS, CUSTOM, or NOSASL

SPARK_USERNAME

Username for LDAP authentication

SPARK_PASSWORDsecret

Password for LDAP authentication

SPARK_KERBEROS_SERVICE_NAME

Kerberos service name (default: hive)

Readme

Spark SQL MCP Server

An MCP server that enables AI assistants to query Spark SQL clusters via the Thrift/HiveServer2 protocol.

Works with any HiveServer2-compatible system: Apache Spark, AWS EMR, Hive, Impala, Presto.

Features

Query Spark SQL — Execute read-only SQL queries against your Spark cluster
Schema Discovery — List databases, tables, and describe table structures
Multiple Auth Methods — NONE, LDAP, NOSASL, CUSTOM, and Kerberos authentication
EMR Compatible — Works with AWS EMR clusters out of the box
Read-Only Enforcement — Only SELECT, SHOW, DESCRIBE, EXPLAIN, and WITH statements are allowed
Safety Defaults — Automatic LIMIT clause on unbounded queries, sanitized error messages

Installation

pip install spark-sql-mcp-server

Or run directly with uvx:

uvx spark-sql-mcp-server

Quick Start

1. Set Environment Variables

View the full README on GitHub

Details

Repository: github.com/aidancorrell/spark-sql-mcp-server
Category: Databases
Package: spark-sql-mcp-server
Transport: STDIO
Auth: Required
Registry: MCP registry, active

Related Databases MCP servers

View all

Mcp Clickhouseinfoinlet-marketplace/marketplaceRead-only ClickHouse for AI agents — analytics queries, table inspection, sizes, health.0stars

OriginChainDBoriginchaindb/mcp-serverConnect an MCP host to an OriginChainDB tenant through four tools: SQL, vector search, full-text search, and schema discovery.0stars

DBHubbytebase/dbhubToken-efficient database MCP server for PostgreSQL, MySQL, MariaDB, SQL Server, Oracle, SQLite3.8Kstars

MCP Alchemy Serverrunekaagaard/mcp-alchemyA MCP (model context protocol) server that gives the LLM access to and knowledge about relational databases like SQLite, Postgresql, MySQL & MariaDB, Oracle, an403stars

Faucetfaucetdb/faucetSQL database to MCP tools and REST API: PostgreSQL, MySQL, SQL Server, Oracle, SQLite, Snowflake.170stars

hypequery ClickHousehypequery/hypequeryGoverned ClickHouse analytics for agents: named datasets and metrics as tools, never raw SQL.104stars