PostgreSQL & Google Sheets
MCP servers for interacting with PostgreSQL databases and Google Sheets.
MCP Server Configuration
This documentation covers the setup and capabilities of multiple MCP (Model Context Protocol) servers for database and cloud service integrations.
1. PostgreSQL MCP Servers Setup
Below is the MCP configuration for connecting to multiple PostgreSQL databases:
{
"mcpServers": {
// MCP server for bi_chatbot database - provides SQL query capabilities
"postgres-bi-chatbot": {
"command": "npx", // Node package executor to run packages without installing
"args": [
"-y", // Auto-confirm prompts, skip interactive installation
"@modelcontextprotocol/server-postgres", // MCP PostgreSQL server package
"postgresql://postgres:cikalmerdeka@localhost:5432/bi_chatbot" // Connection string: protocol://user:password@host:port/database
]
},
// MCP server for airbnb database - provides SQL query capabilities
"postgres-airbnb": {
"command": "npx", // Node package executor to run packages without installing
"args": [
"-y", // Auto-confirm prompts, skip interactive installation
"@modelcontextprotocol/server-postgres", // MCP PostgreSQL server package
"postgresql://postgres:cikalmerdeka@localhost:5432/airbnb" // Connection string: protocol://user:password@host:port/database
]
},
// MCP server for Google Sheets - provides spreadsheet management capabilities
"googlesheets": {
"url": "https://mcp.composio.dev/partner/composio/googlesheets/mcp?customerId=YOUR_CUSTOMER_ID&agent=cursor" // Composio Google Sheets MCP endpoint with OAuth integration
}
}
}
1.1 PostgreSQL MCP Server Setup
The PostgreSQL MCP servers use the official package from the Model Context Protocol:
"postgres-bi-chatbot": {
"command": "npx", // Node package executor to run packages without installing
"args": [
"-y", // Auto-confirm prompts, skip interactive installation
"@modelcontextprotocol/server-postgres", // MCP PostgreSQL server package
"postgresql://postgres:cikalmerdeka@localhost:5432/bi_chatbot" // Connection string: protocol://user:password@host:port/database
]
}
Configuration Parameters
- command: "npx" - Node package executor that runs packages without installing them globally
- args: Array containing the package and connection details
- Connection String Format:
postgresql://username:password@host:port/database
Available Tools
The official PostgreSQL MCP server provides the following capabilities:
Database Schema Exploration
list_schemas
- List all database schemas available in the PostgreSQL instancelist_tables
- List database objects (tables, views, sequences) within a specified schemadescribe_table
- Provides detailed information about a specific table (columns, constraints, indexes)
Query Execution
query
- Execute read-only SQL SELECT queries with optional prepared statement parametersread_query
- Safe read-only query execution with result limiting and timeout controls
Connection Management
- Automatic connection pooling and management
- Secure credential handling through connection strings
- Support for SSL connections and advanced PostgreSQL features
Capabilities & Limitations
✅ What it CAN do:
- Execute read-only SQL queries (SELECT statements)
- Browse database schemas and table structures
- Inspect table columns, data types, and constraints
- View indexes and foreign key relationships
- Support for complex queries with JOINs, aggregations, and subqueries
- Prepared statement parameter binding for security
❌ What it CANNOT do:
- Execute write operations (INSERT, UPDATE, DELETE, DROP)
- Modify database schema or structure
- Create or drop tables, indexes, or other database objects
- Execute stored procedures or functions
- Administrative operations (user management, permissions)
Best for: Safe database exploration, data analysis, reporting, and read-only business intelligence queries.
Advanced Alternative: Postgres MCP Pro
For production environments requiring more advanced capabilities, consider Postgres MCP Pro which includes:
- Database Health Checks - Analyze index health, connection utilization, buffer cache
- Index Tuning - AI-powered index recommendations using optimization algorithms
- Query Performance Analysis - EXPLAIN plans and query optimization suggestions
- Safe SQL Execution - Configurable read-only and restricted modes
- Workload Analysis - Identify slow queries and resource bottlenecks
4. Configuration Explanation
- npx: Node package executor that runs packages without installing them globally
- -y: Auto-confirms prompts and skips interactive installation
- @modelcontextprotocol/server-postgres: The official PostgreSQL MCP server package
- Connection String Format:
postgresql://username:password@host:port/database
2. LangGraph Documentation MCP Server Setup
The LangGraph documentation MCP server provides access to comprehensive documentation for LangGraph and LangChain:
"langgraph-docs-mcp": {
"command": "uvx", // Python package runner that executes packages from PyPI
"args": [
"--from", // Specify the package to install from
"mcpdoc", // Package name containing the MCP documentation server
"mcpdoc", // Command to run from the package
"--urls", // Documentation sources to index
"LangGraph:https://langchain-ai.github.io/langgraph/llms.txt LangChain:https://python.langchain.com/llms.txt", // Space-separated list of documentation URLs
"--transport", // Communication protocol
"stdio" // Standard input/output transport
]
}
Configuration Parameters
- command: "uvx" - Python package runner for executing packages from PyPI without installation
- args: Array containing package specification and configuration options
- --from: Specifies the source package (mcpdoc)
- --urls: Documentation sources in format "Name:URL" separated by spaces
- --transport: Communication protocol (stdio for standard input/output)
Available Tools
The LangGraph documentation MCP server provides the following capabilities:
Documentation Access
list_doc_sources
- List all available documentation sources and their URLsfetch_docs
- Fetch and parse documentation from specific URLs or local files
Content Retrieval
- Access to comprehensive LangGraph documentation covering:
- Memory types (short-term, long-term, semantic, episodic, procedural)
- Agent architectures and multi-agent systems
- Graph construction and execution
- Streaming and persistence capabilities
- Human-in-the-loop workflows
- Platform deployment options
Search and Navigation
- Structured access to documentation sections
- Cross-referencing between related topics
- Real-time documentation updates from official sources
Capabilities & Limitations
✅ What it CAN do:
- Provide detailed explanations of LangGraph concepts and features
- Access the latest documentation from official sources
- Explain implementation patterns and best practices
- Offer code examples and tutorials
- Cover both basic and advanced LangGraph usage
❌ What it CANNOT do:
- Execute or run LangGraph code directly
- Modify or update the documentation sources
- Provide real-time debugging of LangGraph applications
- Access private or custom documentation sources
Best for: Learning LangGraph concepts, understanding implementation patterns, getting detailed explanations of features, and accessing comprehensive documentation for development guidance.
3. Google Sheets MCP Server Setup
The Google Sheets MCP server is provided by Composio and requires OAuth authentication:
"googlesheets": {
"url": "https://mcp.composio.dev/partner/composio/googlesheets/mcp?customerId=YOUR_CUSTOMER_ID&agent=cursor"
}
Configuration Parameters
- url: Composio-hosted MCP server endpoint for Google Sheets integration
- customerId: Your unique Composio customer identifier (obtained during account setup)
- agent: Target application (cursor for Cursor IDE integration)
Available Tools
Core Sheet Operations
GOOGLESHEETS_ADD_SHEET
- Add new worksheets/tabs to spreadsheetGOOGLESHEETS_DELETE_SHEET
- Remove existing sheetsGOOGLESHEETS_GET_SPREADSHEET_INFO
- Get spreadsheet metadata, sheets list, properties
Row/Column Management
GOOGLESHEETS_INSERT_DIMENSION
- Insert rows/columns at specific positionsGOOGLESHEETS_DELETE_DIMENSION
- Delete specified rows/columnsGOOGLESHEETS_APPEND_DIMENSION
- Add rows/columns to the end
Filtering & Sorting
GOOGLESHEETS_SET_BASIC_FILTER
- Apply filters and sorting to data rangesGOOGLESHEETS_CLEAR_BASIC_FILTER
- Remove existing filters
Advanced Features
GOOGLESHEETS_SEARCH_DEVELOPER_METADATA
- Find custom metadata entries
Connection Management
GOOGLESHEETS_CHECK_ACTIVE_CONNECTION
- Verify authentication statusGOOGLESHEETS_INITIATE_CONNECTION
- Start OAuth authentication process
Capabilities & Limitations
✅ What it CAN do:
- Create and delete sheets/worksheets
- Manage sheet structure (rows, columns)
- Set up filters and sorting
- Handle spreadsheet metadata
- OAuth authentication management
❌ What it CANNOT do:
- Read or write cell data/values
- Apply cell formatting (colors, fonts, borders)
- Create charts or pivot tables
- Handle cell formulas
- Batch data operations
Best for: Spreadsheet structure management, template creation, and data organization setup.
5. Connected Services
Databases
- bi_chatbot: Olist e-commerce dataset with 8 tables (customers, orders, products, etc.)
- airbnb: Airbnb dataset for analysis
Cloud Services
- Google Sheets: Spreadsheet management and structure operations
Documentation Services
- LangGraph Documentation: Comprehensive documentation access for LangGraph and LangChain frameworks
6. Usage
After configuration, restart Cursor to enable the MCP servers. Each database will have its own query tool available in the chat interface.
Authentication Requirements
PostgreSQL Servers
- Pre-configured credentials via connection strings in mcp.json
- Automatic connection - No additional setup required after configuration
- Secure connections - Supports SSL and advanced PostgreSQL authentication methods
Google Sheets Server
- OAuth authentication required - One-time browser-based setup
- Browser authorization - Opens redirect URL for Google account access
- Persistent sessions - Authentication persists across MCP client restarts
Related Servers
CData Sage 300
A read-only MCP server by CData that enables LLMs to query live data from Sage 300.
Supabase
Connects to Supabase platform for database, auth, edge functions and more.
Microsoft Fabric Analytics
An analytics server providing tools for interacting with the Microsoft Fabric data platform.
COTI Blockchain MCP
Interact with the COTI blockchain for private token and NFT operations, including ERC20 and ERC721, with privacy features.
IMF Data MCP
Retrieve and process economic data from the International Monetary Fund (IMF) API, including datasets, time series, indicators, and countries.
Tigris Data
A serverless NoSQL database and search platform.
CData Connect Cloud
Query and manage data through CData Connect Cloud, providing a unified interface to various data sources.
MCP-MySQL-Ops
You are working with the MCP MySQL Operations Server, a powerful tool that provides comprehensive MySQL database monitoring and analysis capabilities through natural language queries. This server offers 19 specialized tools for database administration, performance monitoring, and system analysis.
CData Amazon Redshift
Access and manage Amazon Redshift data using the CData MCP Server, which requires an external CData JDBC Driver.
LanceDB
Interact with on-disk documents using agentic RAG and hybrid search via LanceDB.