Databases & Data Stores MCP servers allow AI assistants to query and manage relational databases, document stores, time-series engines, and vector repositories via the Model Context Protocol. Connect tools like Claude Desktop and Cursor to systems such as MongoDB, YugabyteDB, VictoriaMetrics, and ODBC-compliant databases for automated schema inspection, data retrieval, and querying.
Databases & Data Stores MCP servers let AI assistants query, inspect, and update structured and unstructured data repositories directly through the Model Context Protocol. By exposing database clients as standardized tools and resources, these servers enable language models to execute SQL queries, inspect document schemas, retrieve time-series metrics, and pull records from relational engines, document stores, and vector indices.
When evaluating servers in this category, focus on three practical operational criteria:
stdio) or as a remote service over Server-Sent Events (SSE) or HTTP. Local engines like SQLite or self-hosted instances benefit from stdio execution within your client workspace, while distributed clusters require stable network routing.| MCP server | What it connects to | Type | Repository |
|---|---|---|---|
| Zero-Vector v3 | Zero-Vector v3 acts like a sophisticated long-term brain for AI assistants and digital personas. In simple terms, it prevents an | Open Source | https://github.com/fernicar/zero-vector-3 |
| Victoriametrics Mcp Server | Victoriametrics Mcp Server is an MCP server that connects large language model interfaces to VictoriaMetrics time series databases. Built for | Open Source | https://github.com/yincongcyincong/VictoriaMetrics-mcp-server |
| YugabyteDB MCP Server | The YugabyteDB MCP Server acts as a bridge between artificial intelligence and a YugabyteDB database, allowing AI models to "talk" | Open Source | https://github.com/yugabyte/yugabytedb-mcp-server |
| Movies MCP Server | Movies MCP Server is an MCP server that connects Model Context Protocol clients directly to a PostgreSQL database for comprehensive | Unknown | |
| ODBC Server via PyODBC | ODBC Server via PyODBC is an MCP server that provides SQL and metadata inspection tools for databases connected through ODBC | Open Source | https://github.com/OpenLinkSoftware/mcp-pyodbc-server |
| MongoDB MCP Server | MongoDB MCP Server is an MCP server that provides tools and prompts for interacting directly with a MongoDB database. Built | Unknown | https://github.com/jb-7-jay/learn-mcp-server-mongo |
| Salesforce Einstein by CData | Salesforce Einstein by CData is an MCP server that exposes Salesforce Einstein data as relational SQL models for AI assistants. | Unknown | https://github.com/CDataSoftware/salesforce-einstein-mcp-server-by-cdata |
| RewindDB | RewindDB is an MCP server that interfaces with the local Rewind.ai SQLite database to expose recorded audio transcripts and optical | Unknown | https://github.com/pedramamini/RewindMCP |
| YouTube Data MCP | The YouTube Data MCP acts as a smart bridge between AI models and the vast library of content on YouTube. | Open Source | https://github.com/kirbah/mcp-youtube |
| Zero-Vector MCP | Zero-Vector MCP acts like a high-powered digital brain for AI assistants, giving them the ability to remember past conversations and | Unknown | https://github.com/MushroomFleet/zero-vector-MCP |
The right option depends on your backend architecture. For relational data requiring ODBC drivers, ODBC Server via PyODBC connects smoothly using standard local configuration. If you work with document databases, MongoDB MCP Server exposes document querying tools directly inside Claude Desktop. Teams monitoring operational metrics often use Victoriametrics Mcp Server to pull time-series data into prompts.
Add the chosen server command and environment variables to your Cursor configuration file. For instance, point the configuration to your Python or Node runtime executing MongoDB MCP Server, YugabyteDB MCP Server, or OceanBase. Ensure database credentials and connection URIs are securely passed via the configuration environment block so the model can call database tools during code generation.
Yes. XiYan MCP Server is specifically built to enable natural language queries to databases using XiyanSQL. Additionally, servers like ODBC Server via PyODBC, OceanBase, and Movies MCP Server provide schema inspection tools that allow models to inspect table definitions, construct valid SQL statements, and retrieve query results safely.
Yes. Several servers provide restricted read access to protect production data. For example, Salesforce Einstein by CData functions as a dedicated read-only MCP server for querying live analytics data. Similarly, RewindDB provides read access to local Rewind.ai SQLite transcripts and OCR data without modifying underlying records.