Postgres MCP Server
# Postgres MCP Server
MCP server for PostgreSQL database management and operations, built with a sophisticated enterprise-grade architecture.
## Quick Setup
### 1. Installation
```bash
npm install
npm run build
```
### 2. Claude Desktop Configuration
Add this to your Claude Desktop `claude_desktop_config.json`:
**Windows:**
```json
{
"mcpServers": {
"postgres": {
"command": "node",
"args": ["C:\\path\\to\\postgres-mcp\\dist\\index.js"],
"env": {
"DATABASE_URL": "postgresql://username:password@localhost:5432/dbname"
}
}
}
}
```
**macOS/Linux:**
```json
{
"mcpServers": {
"postgres": {
"command": "node",
"args": ["/path/to/postgres-mcp/dist/index.js"],
"env": {
"DATABASE_URL": "postgresql://username:password@localhost:5432/dbname"
}
}
}
}
```
### 3. Environment Configuration
**Option A: Via Claude Desktop config (recommended)**
```json
{
"mcpServers": {
"postgres": {
"command": "node",
"args": ["/Users/itsalfredakku/McpServers/postgres-mcp/dist/index.js"],
"env": {
"DATABASE_URL": "postgresql://postgres:password@localhost:5432/mydb",
"POOL_MAX": "20",
"LOG_LEVEL": "info"
}
}
}
}
```
**Option B: Using .env file**
Create `.env` in the project root:
```env
DATABASE_URL=postgresql://username:password@localhost:5432/dbname
POOL_MAX=10
LOG_LEVEL=info
```
## Features
- **Database Operations**: Query, insert, update, delete operations
- **Schema Management**: Create, alter, drop tables and indexes
- **Transaction Management**: Begin, commit, rollback transactions
- **Connection Management**: Advanced connection pooling
- **Data Management**: Import/export, backup/restore operations
- **Monitoring**: Performance metrics and query analysis
- **Admin Operations**: User management, permissions, database administration
## Installation
```bash
npm install
```
## Configuration Options
### Database Connection
```env
# Required - Primary connection string
DATABASE_URL=postgresql://username:password@localhost:5432/dbname
# Alternative - Individual connection parameters
POSTGRES_HOST=localhost
POSTGRES_PORT=5432
POSTGRES_USER=postgres
POSTGRES_PASSWORD=your_password
POSTGRES_DATABASE=your_database
POSTGRES_SSL=false
```
### Connection Pool Settings
```env
POOL_MIN=2 # Minimum connections
POOL_MAX=10 # Maximum connections
POOL_IDLE_TIMEOUT=30000 # Idle timeout (ms)
POOL_ACQUIRE_TIMEOUT=60000 # Acquire timeout (ms)
```
### Performance & Caching
```env
CACHE_ENABLED=true # Enable query result caching
CACHE_TTL=300000 # Cache TTL (ms)
LOG_LEVEL=info # Logging level (error|warn|info|debug)
SQL_LOGGING=false # Log SQL queries
```
## Usage
### Development
```bash
npm run dev
```
### Production
```bash
npm run build
npm start
```
### Testing
```bash
npm run test
npm run test:queries
```
## Tools
### Database Operations
- `query` - Execute SQL queries with transaction support, explain plans, analysis
- `tables` - List, create, alter, drop tables with detailed metadata
- `schemas` - **FULLY IMPLEMENTED** Create, drop, list schemas and manage permissions
- `indexes` - **FULLY IMPLEMENTED** Create, drop, analyze, reindex with usage statistics
### Data Management
- `data` - Insert, update, delete operations with bulk support
- `transactions` - Begin, commit, rollback with savepoint support
### Administration & Security
- `admin` - **FULLY IMPLEMENTED** Complete database administration and maintenance
- `permissions` - Complete user/role/privilege management
- `security` - SSL, authentication, encryption, auditing
- `monitoring` - Performance metrics and analysis
- `connections` - Connection pool management
### Schema Management Features ✅
- **Schema Operations**: Create, drop, list all schemas
- **Permission Management**: View and manage schema-level permissions
- **Owner Management**: Set schema ownership during creation
- **Conditional Operations**: IF EXISTS, IF NOT EXISTS support
- **System Schema Filtering**: Distinguish between user and system schemas
### Index Management Features ✅
- **Index Operations**: Create, drop, list, reindex indexes
- **Performance Analysis**: Analyze index usage statistics
- **Unused Index Detection**: Find indexes that are never used
- **Multiple Index Types**: Support for btree, hash, gist, gin, brin
- **Concurrent Operations**: Create and reindex with CONCURRENTLY
- **Size Monitoring**: Index size tracking and reporting
### Database Administration Features ✅
- **Database Information**: Complete database stats and configuration
- **User Management**: Create, drop, list users with detailed privileges
- **Permission Control**: Grant/revoke permissions on tables and schemas
- **Maintenance Operations**: VACUUM, ANALYZE, REINDEX with options
- **System Monitoring**: Connection counts, database size, uptime tracking
- **Configuration Access**: View database settings and parameters
## Architecture
The server follows a modular architecture with:
- **Configuration Management** - Environment and file-based configuration
- **Connection Pooling** - Advanced PostgreSQL connection management
- **Domain APIs** - Separated concerns for different database operations
- **Validation** - Comprehensive parameter validation
- **Error Handling** - Robust error handling with retries
- **Caching** - Intelligent caching for performance
- **Logging** - Structured logging with Winston
## Troubleshooting
### Common Issues
**Connection Refused**
```bash
# Check if PostgreSQL is running
brew services list | grep postgresql
# or
sudo systemctl status postgresql
# Test connection manually
psql -h localhost -p 5432 -U postgres -d your_database
```
**Permission Denied**
```sql
-- Grant necessary permissions
GRANT CONNECT ON DATABASE your_database TO your_user;
GRANT USAGE ON SCHEMA public TO your_user;
GRANT CREATE ON SCHEMA public TO your_user;
```
**MCP Server Not Found**
- Ensure the path in `claude_desktop_config.json` is absolute
- Verify `npm run build` completed successfully
- Check that `dist/index.js` exists
### Debug Mode
Set environment variables for detailed logging:
```json
{
"mcpServers": {
"postgres": {
"command": "node",
"args": ["/path/to/postgres-mcp/dist/index.js"],
"env": {
"DATABASE_URL": "postgresql://user:pass@localhost:5432/db",
"LOG_LEVEL": "debug",
"SQL_LOGGING": "true"
}
}
}
}
```
## Database Permissions Setup
### Full Admin Access
For complete database management capabilities, ensure your PostgreSQL user has appropriate privileges:
```sql
-- Connect as superuser (postgres)
psql -U postgres
-- Create a dedicated MCP user with admin privileges
CREATE USER mcp_admin WITH PASSWORD 'secure_password';
ALTER USER mcp_admin SUPERUSER;
ALTER USER mcp_admin CREATEDB;
ALTER USER mcp_admin CREATEROLE;
ALTER USER mcp_admin REPLICATION;
-- Or grant specific privileges without superuser
CREATE USER mcp_user WITH PASSWORD 'secure_password';
GRANT ALL PRIVILEGES ON DATABASE your_database TO mcp_user;
GRANT ALL PRIVILEGES ON ALL TABLES IN SCHEMA public TO mcp_user;
GRANT ALL PRIVILEGES ON ALL SEQUENCES IN SCHEMA public TO mcp_user;
GRANT ALL PRIVILEGES ON ALL FUNCTIONS IN SCHEMA public TO mcp_user;
-- Grant schema usage and creation
GRANT USAGE, CREATE ON SCHEMA public TO mcp_user;
-- Allow user management (requires elevated privileges)
ALTER USER mcp_user CREATEROLE;
```
### Using MCP Permission Tools
Once connected, you can use the MCP server to manage permissions:
```typescript
// List all users and their privileges
await mcpServer.callTool('permissions', { operation: 'list_users' });
// Create a new user
await mcpServer.callTool('permissions', {
operation: 'create_user',
username: 'newuser',
password: 'password123',
attributes: { createdb: true, login: true }
});
// Grant all privileges to a user
await mcpServer.callTool('permissions', {
operation: 'grant_all_privileges',
username: 'newuser',
database: 'mydatabase'
});
// Check user permissions
await mcpServer.callTool('permissions', {
operation: 'check_permissions',
username: 'newuser'
});
```
## License
MITTDQS
Scored across 11 tools
Most tools have distinct purposes targeting specific database management areas like connections, data, indexes, and schemas. However, there is some overlap between 'admin' and 'permissions' (both mention user/role management) and 'admin' and 'security' (both cover authentication aspects), which could cause minor confusion for an agent.
All tool names follow a consistent singular noun pattern (e.g., admin, connections, data, indexes) without mixing conventions. This predictability makes it easy for an agent to understand and navigate the tool set.
With 11 tools, this server is well-scoped for comprehensive Postgres database management, covering administration, operations, monitoring, and security. Each tool earns its place by addressing a distinct aspect of the domain.
The tool set provides complete coverage for Postgres management, including CRUD operations (data), schema and table management, indexing, monitoring, security, and transaction support. There are no obvious gaps, enabling agents to handle full database workflows.