AI
MCP

Best MCP Servers for Database Management of 2025

Author: Antonello Zanini
Length: 12 min
Type: Guide
Published: 2025-10-20

intro

Let’s explore the most widely used and powerful MCP servers for database management that you can access directly through your AI agents.

Imagine managing your database using simple prompts in natural language. What seemed like sci-fi just a few years ago is now a reality thanks to technologies that integrate LLMs with databases to convert natural language prompts into precise database queries. This is not magic, but the promise delivered by MCP servers.

With Google recently open-sourcing its own MCP server used internally for database management, the trend of using MCP servers for handling databases is gaining traction and becoming more popular than ever.

In this article, we will showcase the hottest and most useful MCP servers for database management of the year.

The Revolution of MCP in Database Management

MCP (Model Context Protocol) is an open standard developed by Anthropic that enables standardized communication between LLMs and third-party tools. At its core, MCP provides a structured way for LLMs to access, understand, and interact with external services, solutions, and data sources (e.g., local file systems and databases).

When it comes to database integration, an MCP server acts as an intelligent intermediary between the database server and the LLM (typically within an AI agent system). In other words, it translates natural language requests from an LLM into structured queries or commands that the database can execute.

An MCP database server in action
An MCP database server in action

This gives LLMs the ability not only to retrieve specific data but also to comprehend database schemas, run complex queries, and manage database structures. The benefits are significant, as AI agents can to perform real-time data analysis, automate report generation, and handle intelligent database administration. All that, from simple prompts in natural language!

An MCP server for database management also opens the door for non-technical users to query, modify, create, and populate new database schemas with simple instructions.

Aspects to Consider When Evaluating MCP Servers for Databases

The Awesome MCP Servers repository alone lists over 40 MCP servers for handling or connecting to databases. With such a long list of open-source solutions, it can be difficult to decide which is best for your needs.

To evaluate these MCP database tools, you should focus on the aspects summarized in the following table:

AspectDescription
Developed byWhether the tool is official and developed by a known company (or the company behind the database), or by the community.
Supported databasesThe list of databases the MCP server can interact with.
ToolsThe tools/capabilities provided by the MCP server.
Programming languageThe programming language the MCP server is built with, which helps you understand whether it can be easily integrated into your tech stack.
GitHub starsThe number of stars on GitHub, indicating popularity and community interest.

Top Open-Source Database MCP Server?

Discover the best MCP servers for database management, selected based on the criteria presented earlier.

Note: MCP is still a fresh protocol, and many new servers are released every day. Thus, this list only provides a snapshot of the available MCP servers for database management at the time of writing. Also, keep in mind that this is simply a list, not a ranking.

googleapis/genai-toolbox

At Google Cloud Next 25, Google announced that it would open-source what was originally called Gen AI Toolbox for Databases.

Now known as MCP Toolbox for Databases, googleapis/genai-toolbox is an open-source MCP server that simplifies building tools by managing complexities like connection pooling, authentication, and more.

For comprehensive details, refer to the full documentation.

Developed by: Google

Supported databases: PostgreSQL, MySQL, SQL Server, Neo4j, Dgraph, Spanner, BigQuery, Cloud SQL (both MySQL and PostgreSQL), AlloyDB, and more

Tools: A long list of database-specific tools that, as a whole, offers:

  • Ask complex questions in plain English from your IDE, with no SQL needed.
  • Describe your needs and let AI handle queries, table creation, indexing, and more.
  • Let AI generate app code and tests using your live database schema for faster development.
  • Cut down time spent on setup, boilerplate, and error-prone schema migrations.

Programming language: Go

GitHub stars: 5.3k+

supabase-community/supabase-mcp

supabase-community/supabase-mcp standardizes how LLMs interact with Supabase, a PostgreSQL-compatible database. It connects AI assistants directly to your Supabase project, enabling them to manage tables, retrieve configuration, and run data queries.

Find out more in the official documentation.

Developed by: The Supabase community

Supported databases: Supabase.

Tools:

  • list_projects: Lists all Supabase projects linked to the user.
  • get_project: Retrieves details for a specific project.
  • create_project: Creates a brand-new Supabase project.
  • pause_project: Temporarily suspends a project.
  • restore_project: Reactivates a paused project.
  • list_organizations: Lists all organizations the user belongs to.
  • get_organization: Fetches information about a particular organization.
  • get_cost: Estimates the cost of creating a new project or branch within an organization.
  • confirm_cost: Confirms the user understands costs before creating a new project or branch.
  • search_docs: Searches Supabase documentation for the latest guides and references.
  • list_tables: Shows all tables in the specified schemas.
  • list_extensions: Lists all extensions installed in the database.
  • list_migrations: Lists migration history for the database.
  • apply_migration: Applies a SQL migration, tracking schema changes.
  • execute_sql: Runs raw SQL queries that don’t alter the schema.
  • get_logs: Retrieves logs for a Supabase project by service type to aid in debugging.
  • get_advisors: Fetches advisory notices highlighting security or performance issues.
  • get_project_url: Gets the API URL for a given project.
  • get_anon_key: Retrieves the anonymous API key for client access.
  • generate_typescript_types: Generates TypeScript types based on your database schema.
  • list_edge_functions: Lists all Edge Functions deployed in a project.
  • deploy_edge_function: Deploys or updates an Edge Function within a project.
  • create_branch: Creates a development branch with migrations copied from production.
  • list_branches: Lists all active development branches.
  • delete_branch: Removes a development branch.
  • merge_branch: Merges migrations and functions from a dev branch into production.
  • reset_branch: Rolls back migrations in a dev branch to an earlier state.
  • rebase_branch: Re-applies a dev branch on top of production to resolve drift.
  • list_storage_buckets: Lists all storage buckets in a project.
  • get_storage_config: Retrieves current storage configuration.
  • update_storage_config: Updates storage settings for a project (paid feature).

Programming language: TypeScript

GitHub stars: 1.8k+

chroma-core/chroma-mcp

chroma-core/chroma-mcp is an MCP server that leverages Chroma to provide advanced data retrieval. It allows AI models to create collections from generated data and user inputs, and retrieve them using vector search, full-text search, metadata filters, and more.

Learn more on the official docs.

Developed by: Chroma team (the company behind the Chroma vector database)

Supported databases: Chroma

Tools:

  • chroma_list_collections: Lists all collections, with support for pagination.
  • chroma_create_collection: Creates a new collection, optionally configuring HNSW parameters.
  • chroma_peek_collection: Displays a sample of documents contained in a collection.
  • chroma_get_collection_info: Retrieves detailed metadata and settings for a collection.
  • chroma_get_collection_count: Returns the total number of documents within a collection.
  • chroma_modify_collection: Updates a collection’s name or associated metadata.
  • chroma_delete_collection: Removes an entire collection from the database.
  • chroma_add_documents: Adds documents, allowing optional metadata and custom IDs.
  • chroma_query_documents: Performs semantic search queries with advanced filtering options.
  • chroma_get_documents: Fetches documents by IDs or filter criteria, with pagination support.
  • chroma_update_documents: Modifies existing documents’ content, metadata, or embeddings.
  • chroma_delete_documents: Deletes specific documents from a collection.

Programming language: Python

GitHub stars: 226+

mongodb-js/mongodb-mcp-server

mongodb-js/mongodb-mcp-server is an official MCP server for interacting with MongoDB databases, including MongoDB Atlas.

Developed by: MongoDB

Supported databases: MongoDB

Tools:

  • atlas-list-orgs: Lists MongoDB Atlas organizations.
  • atlas-list-projects: Lists projects within MongoDB Atlas.
  • atlas-create-project: Creates a new project in MongoDB Atlas.
  • atlas-list-clusters: Lists clusters in MongoDB Atlas.
  • atlas-inspect-cluster: Inspects details of a specific MongoDB Atlas cluster.
  • atlas-create-free-cluster: Creates a free-tier MongoDB Atlas cluster.
  • atlas-connect-cluster: Connects to a MongoDB Atlas cluster.
  • atlas-inspect-access-list: Views IP/CIDR access lists for MongoDB Atlas clusters.
  • atlas-create-access-list: Configures IP/CIDR access lists for clusters.
  • atlas-list-db-users: Lists database users in MongoDB Atlas.
  • atlas-create-db-user: Creates a new MongoDB Atlas database user.
  • atlas-list-alerts: Lists alerts for a MongoDB Atlas project.
  • connect: Connects to a MongoDB instance.
  • find: Executes a find query on a MongoDB collection.
  • aggregate: Runs aggregation pipelines on a collection.
  • count: Counts documents in a collection.
  • insert-one: Inserts a single document into a collection.
  • insert-many: Inserts multiple documents into a collection.
  • create-index: Creates an index on a collection.
  • update-one: Updates a single document in a collection.
  • update-many: Updates multiple documents in a collection.
  • rename-collection: Renames a MongoDB collection.
  • delete-one: Deletes a single document from a collection.
  • delete-many: Deletes multiple documents from a collection.
  • drop-collection: Drops a collection from the database.
  • drop-database: Deletes an entire MongoDB database.
  • list-databases: Lists all databases accessible by the connection.
  • list-collections: Lists all collections within a specified database.
  • collection-indexes: Describes indexes on a collection.
  • collection-schema: Describes the schema of a collection.
  • collection-storage-size: Returns the size of a collection in megabytes.
  • db-stats: Provides statistics about a MongoDB database.

Programming language: TypeScript

GitHub stars: 399+

ClickHouse/mcp-clickhouse

ClickHouse/mcp-clickhouse is an MCP server for ClickHouse, a fast, resource-efficient, real-time data warehouse and open-source database. This integration enables natural language queries and automated database management directly via LLMs.

Developed by: ClickHouse team

Supported databases: ClickHouse, chDB

Tools:

  • run_select_query: Executes SQL queries on your ClickHouse cluster.
  • list_databases: Lists all databases available on your ClickHouse cluster.
  • list_tables: Lists all tables within a specified database.
  • run_chdb_select_query: Executes SQL queries using chDB’s embedded OLAP engine.

Programming language: Python

GitHub stars: 438 stars

isaacwasserman/mcp-snowflake-server

isaacwasserman/mcp-snowflake-server is a community-developed MCP server that facilitates database interaction with Snowflake, a cloud data platform offering data warehouse-as-a-service. It supports running SQL queries, exposes data insights, and provides schema context—enabling AI-driven exploration and analysis within Snowflake environments.

Developed by: Community

Supported databases: Snowflake.

Tools:

  • read_query: Runs SELECT statements to fetch data.
  • write_query: Executes INSERT, UPDATE, or DELETE operations.
  • create_table: Generates new tables in the database.
  • list_databases: Retrieves all databases in the Snowflake instance.
  • list_schemas: Shows all schemas under a given database.
  • list_tables: Lists tables within a specified database and schema.
  • describe_table: Displays column details for a particular table.
  • append_insight: Attaches new data insights to a memo resource.

Programming language: Python

GitHub stars: 128+

prisma/mcp

prisma/mcp is an MCP server built to integrate your AI agent with the Prisma Postgres service. That is a managed PostgreSQL database offering always-on availability and pay-as-you-go pricing.

As explained in the official docs, note that a remote Prisma MCP server is also available.

Developed by: Prisma

Supported databases: PostgreSQL

Tools:

  • CreateBackupTool: Generates a new managed backup for a Prisma Postgres database.
  • CreateConnectionStringTool: Generates a new connection string for a Prisma Postgres database using the specified ID.
  • CreateRecoveryTool: Restores a Prisma Postgres database to a new instance using the provided backup ID.
  • DeleteConnectionStringTool: Removes a connection string identified by the given connection string ID.
  • DeleteDatabaseTool: Deletes a Prisma Postgres database identified by the specified ID.
  • ListBackupsTool: Retrieves a list of available Prisma Postgres backups for a given database and environment ID.
  • ListConnectionStringsTool: Retrieves all connection strings available for a Prisma Postgres database by database and environment ID.
  • ListDatabasesTool: Lists all Prisma Postgres databases accessible within the user's workspace.
  • ExecuteSqlQueryTool: Runs a SQL query on a Prisma Postgres database specified by its ID.
  • IntrospectSchemaTool: Examines the schema of a Prisma Postgres database with the given ID.

Programming language: JavaScript

GitHub stars: 18+

singlestore-labs/mcp-server-singlestore

The singlestore-labs/mcp-server-singlestore MCP database server enables Claude Desktop, Cursor, and other compatible MCP clients to interact with SingleStore using natural language. Its ultimate goal is to simplify complex database operations within SingleStore, a distributed and relational SQL database.

Developed by: SingleStore

Supported databases: SingleStore

Tools:

  • workspace_groups_info: Fetches information about accessible workspace groups.
  • workspaces_info: Retrieves details of workspaces within a given group.
  • organization_info: Provides information about the user’s current organization.
  • list_of_regions: Returns a list of regions where workspaces are supported.
  • execute_sql: Performs SQL operations on a connected workspace.
  • list_virtual_workspaces: Lists all available starter workspaces.
  • create_virtual_workspace: Creates a new starter workspace instance.
  • execute_sql_on_virtual_workspace: Runs SQL queries on a virtual workspace.
  • list_notebook_samples: Retrieves available notebook samples in SingleStore Spaces.
  • list_personal_files: Retrieves all files stored in the user’s personal space.
  • create_scheduled_job: Sets up a new scheduled job to run a notebook.
  • get_job_details: Fetches details of a specific scheduled job.
  • list_job_executions: Shows the execution history for a given job.

Programming language: Python.

GitHub stars: 22+

LucasHild/mcp-server-bigquery

LucasHild/mcp-server-bigquery is an MCP database server that connects to Google’s BigQuery, a fully managed, serverless, SQL-based data warehouse. It allows LLMs to explore database schemas and run queries seamlessly.

Developed by: Community

Supported databases: BigQuery

Tools:

  • execute-query: Runs SQL statements using BigQuery’s dialect.
  • list-tables: Retrieves all tables within the given BigQuery project and datasets.
  • describe-table: Fetches the schema details of a specified BigQuery table.

Programming language: Python

GitHub stars: 107+

esigncomputer/mysql_mcp_server

designcomputer/mysql_mcp_server supports communication between AI applications and MySQL. It leads to safer and more structured database exploration and analysis through a controlled interface.

Developed by: Community

Supported databases: MySQL

Tools: A set of MySQL-dedicated tools to:

  • List accessible MySQL tables as resources
  • Retrieve contents from tables
  • Run SQL queries with robust error management
  • Ensure secure database access via environment variables
  • Detailed logging and monitoring

Programming language: Python

GitHub stars: 689+

Best MCP Servers for Database Management: Summary Table

Use the summary table below to compare the MCP database servers featured in this article:

MCP ServerDeveloped bySupported databasesProgramming languageGitHub stars
googleapis/genai-toolboxGooglePostgreSQL, MySQL, SQL Server, Neo4j, Dgraph, Spanner, BigQuery, Cloud SQL, AlloyDB, and moreGo5.3k+
supabase-community/supabase-mcpSupabase communitySupabaseTypeScript1.8k+
chroma-core/chroma-mcpChroma teamChromaPython226+
mongodb-js/mongodb-mcp-serverMongoDBMongoDBTypeScript399+
ClickHouse/mcp-clickhouseClickHouse teamClickHouse, chDBPython438
isaacwasserman/mcp-snowflake-serverCommunitySnowflakePython128+
prisma/mcpPrismaPostgreSQLJavaScript18+
singlestore-labs/mcp-server-singlestoreSingleStoreSingleStorePython22+
LucasHild/mcp-server-bigqueryCommunityBigQueryPython107+
designcomputer/mysql_mcp_serverCommunityMySQLPython689+

Conclusion

In this blog post, we explored some of the best MCP servers for database integration available at the time of writing. We highlighted the top open-source servers and summarized their main tools and aspects relevant to supporting AI integration with databases.

Dbvis download link img
About the author
Antonello Zanini

Antonello is a software engineer, and often refers to himself as a technology bishop. His mission is to spread knowledge through writing.

The Table Icon
Sign up to receive The Table's roundup
More from the table
Title Author Tags Length Published
title

Best Databases for Agentic RAG Scenarios

 author Antonello Zanini tags AI Recommendations 8 min 2025-10-15
title

pgvectorscale: An Extension for Improved Vector Search in Postgres

 author Antonello Zanini tags AI POSTGRESQL Vectors 9 min 2025-09-03
title

SQL Server Vector Data Type, Search, and Indexing

 author Antonello Zanini tags AI SQL SERVER Vectors 8 min 2025-08-25
title

Oracle 23ai: What’s New? Everything You Need to Know at a Glance

 author Antonello Zanini tags AI ORACLE 7 min 2025-08-04
title

MySQL LOCATE Function: Find Substring Position

 author Antonello Zanini tags MySQL 7 min 2025-10-22
title

Parsing and SQL Data Types: A Complete Guide

 author Lukas Vileikis tags MySQL SQL 6 min 2025-10-21
title

Top Database CI/CD and Schema Change Tools in 2025

 author Antonello Zanini tags CI/CD Recommendations Schema change 11 min 2025-10-14
title

Best Database Tools for Administrators: Ultimate List

 author TheTable tags DBA PERFORMANCE SECURITY 7 min 2025-10-13
title

Best SQL Clients for Developers: Complete List

 author Antonello Zanini tags Database clients SQL 15 min 2025-10-08
title

Best Database Tools for Business Users: Complete List

 author TheTable tags BI SQL 7 min 2025-10-07
Read more from theTable

The content provided on dbvis.com/thetable, including but not limited to code and examples, is intended for educational and informational purposes only. We do not make any warranties or representations of any kind. Read more here.