DBeast PostgreSQL MCP MCP Server

<div align="center"> <img src="assets/logo.svg" alt="DBeast" width="720" />

A PostgreSQL MCP server that gives AI assistants expert DBA capabilities.

Quick Start · Demo · Tools · Safety · Configuration · Docs

</div>

---

DBeast connects AI assistants such as Claude, Cursor, Windsurf, and VS Code Copilot to PostgreSQL through the Model Context Protocol. Instead of exposing one broad execute_sql escape hatch, DBeast provides 21 focused tools for schema discovery, safe query execution, impact analysis, performance review, security checks, maintenance reporting, replication monitoring, and data quality inspection.

---

Demo

Watch Claude use DBeast MCP tools to audit a PostgreSQL database, identify security and maintenance risks, and preview cleanup impact without executing destructive SQL.

<div align="center">

Watch the demo on YouTube

</div>

---

How It Works

AI assistant  --MCP stdio-->  DBeast server  --asyncpg-->  PostgreSQL
Claude/Cursor                  Python local                 Local, RDS,
Windsurf/VS Code               subprocess                   Supabase, Neon

DBeast runs as a local stdio MCP server. Your IDE or desktop assistant starts it as a subprocess and passes database credentials through environment variables. The assistant calls DBeast tools, DBeast queries PostgreSQL, and structured results come back to the assistant. No HTTP service or extra infrastructure is required.

---

Quick Start

1. Install

git clone https://github.com/snss10/DBeast.git
cd DBeast
pip install -e .

For development:

pip install -e ".[dev]"

Optional: copy .env.example to .env and set your database credentials.

2. Verify

dbeast

Or run the source entry point directly:

python src/server.py

3. Configure Your MCP Client

Minimal Cursor or Windsurf config:

{
  "mcpServers": {
    "dbeast": {
      "type": "stdio",
      "command": "python",
      "args": ["/absolute/path/to/DBeast/src/server.py"],
      "env": {
        "DATABASE_URL": "postgresql://user:password@localhost:5432/mydb"
      }
    }
  }
}

Common config locations:

Client	Config location
Cursor	.mcp.json in project root, or ~/.cursor/.mcp.json globally
VS Code	.vscode/settings.json or user settings with key mcp.servers
Claude Desktop on macOS	~/Library/Application Support/Claude/claude_desktop_config.json
Claude Desktop on Windows	%APPDATA%\Claude\claude_desktop_config.json
Windsurf	.mcp.json

See SETUP.md for full client examples, Docker, RDS, Supabase, Neon, SSH tunnels, AWS Secrets Manager, and troubleshooting.

4. Ask Simple or Complex Questions

Once connected, your assistant can answer quick lookup questions and also run multi-step database investigations.

Simple examples:

Show me the schema for the orders table.
Which queries are slowest right now?
Run a security audit on the public schema.
Generate a Mermaid ERD for the sales schema.

More complex examples:

Before I archive old sessions, estimate how many rows would be affected, identify related tables, and tell me the rollback risk.
Investigate why the dashboard query is slow, explain the execution plan, and suggest safe indexes.
Review the public schema for maintenance issues, security risks, and data quality problems, then summarize the top priorities.
Compare table growth, dead tuples, and index health across all schemas and recommend what to vacuum or reindex first.

---

Tools

DBeast exposes 21 MCP tools across 10 categories.

Connection

Tool	Description
connect	Connect to PostgreSQL, check current status, or discover local databases
disconnect	Close the current database connection
health_check	Verify connectivity, pool health, PostgreSQL version, and extensions

Schema Discovery

Tool	Description
get_schema	List schemas, tables, columns, indexes, relationships, and optional Mermaid ERDs
dependency_analysis	Map object dependencies before renaming, dropping, or changing database objects

Data Access

Tool	Description
execute_query	Run read-only SELECT queries with automatic row-limit injection

Query Analysis

Tool	Description
analyze_query	Parse and inspect query structure, warnings, and optimization hints
query_optimizer	Recommend indexes and rewrites for a given query
analyze_impact	Preview write-query impact, risk level, affected rows, and rollback context without executing

Database Health

Tool	Description
database_health	Review cache hit rates, connections, transaction age, table health, and overall health signals
query_performance	Report slow or expensive queries from PostgreSQL statistics

Security

Tool	Description
security_audit	Inspect roles, privileges, superuser accounts, and public schema exposure
sensitive_data_scan	Detect likely PII or secrets by column names and schema patterns

Maintenance

Tool	Description
maintenance_analysis	Review vacuum status, dead tuples, analyze timestamps, and index health
partition_analysis	Inspect partition health, row distribution, and missing partition risks

Data Quality

Tool	Description
data_quality_report	Analyze null rates, cardinality, value distributions, and outliers
duplicate_detection	Find duplicate rows across selected key columns

Server Config

Tool	Description
configuration_review	Review PostgreSQL configuration and tuning opportunities
replication_status	Inspect replication lag, WAL sender/receiver state, and replication slots

Audit

Tool	Description
get_audit_logs	Retrieve logged MCP tool calls for a given date
list_audit_files	List available audit log files

---

Recommended Workflow

Start by discovering schemas:

get_schema()
get_schema(schema='public')

Run safe read queries:

execute_query(query='SELECT * FROM orders ORDER BY created_at DESC')

Preview risky writes:

analyze_impact(query='DELETE FROM sessions WHERE last_active < now() - interval ''30 days''')

Check health and maintenance:

database_health()
maintenance_analysis(schema='public')
query_performance()

Most analysis tools accept a schema parameter:

maintenance_analysis(schema='public')  -> analyze one schema
maintenance_analysis(schema='all')     -> analyze every schema
get_schema(format='mermaid')           -> generate an ERD diagram

---

Supported Databases

Provider	Connection method
Local PostgreSQL	DATABASE_URL or individual DB_* variables
Docker PostgreSQL	Explicit variables or connect(discover=true)
AWS RDS / Aurora	Direct URL, SSH tunnel, or AWS Secrets Manager
Supabase	Pooler connection string from Dashboard settings
Neon	Connection string from Console connection details
Railway / Render / Fly.io	Provider connection string
Any PostgreSQL host	Standard PostgreSQL URL

---

Configuration

Choose one connection method.

# Full URL
DATABASE_URL=postgresql://user:pass@host:5432/db

# Or individual variables
DB_HOST=localhost
DB_PORT=5432
DB_USER=postgres
DB_PASSWORD=secret
DB_NAME=mydb
DB_SSLMODE=prefer

# Or AWS Secrets Manager
AWS_SECRET_NAME=my-rds-secret
AWS_REGION=us-west-2

You can also connect at runtime:

connect(url='postgresql://user:pass@host:5432/db')
connect(host='localhost', user='postgres', password='secret', database='mydb')
connect(aws_secret_name='my-secret', aws_region='us-west-2')

Key settings:

Variable	Default	Description
DBEAST_DEFAULT_ROW_LIMIT	100	Max rows returned by execute_query
DBEAST_QUERY_TIMEOUT	300	Query execution timeout in seconds
DBEAST_COMMAND_TIMEOUT	300	SQL command timeout in seconds
DBEAST_SSL_VERIFY	true	Set false for SSH tunnels where certificates do not match localhost
DBEAST_SCHEMA_CACHE_TTL	60	Schema cache TTL in seconds, 0 disables caching
DBEAST_AUDIT_ENABLED	true	Log MCP tool calls
DBEAST_AUDIT_DIR	logs/mcp_audit	Audit log directory

See SETUP.md for the complete configuration reference.

---

Safety Model

Query type	What DBeast does
SELECT	Executes with automatic row limits
INSERT / UPDATE / DELETE	Never executed; returns an impact preview
DROP / TRUNCATE	Never executed; reports affected objects and risk

Formatted and JSON responses use a consistent wrapper:

{
  "success": true,
  "data": { "...": "..." },
  "meta": {
    "connected": true,
    "source": "tool"
  }
}

---

Audit Logging

DBeast logs MCP tool calls for accountability and debugging.

DBEAST_AUDIT_ENABLED=true
DBEAST_AUDIT_DIR=logs/mcp_audit

Audit files are stored as daily markdown files and include timestamps, tool names, durations, masked parameters, truncated responses, and errors.

---

Development

pip install -e ".[dev]"
pre-commit install
pytest tests/ -v
ruff check src/ tests/
ruff format src/ tests/

Start the optional local PostgreSQL test database:

docker compose up -d postgres

Legacy Compose:

docker-compose up -d postgres

---

Documentation

• SETUP.md - Full client setup, connection scenarios, configuration, and troubleshooting
• CONTRIBUTING.md - Development setup, tests, style, commits, and PR process
• CODE_OF_CONDUCT.md - Community guidelines

---

License

MIT