postgres-mcp
MCP server for PostgreSQL that enables schema exploration, safe SQL execution, EXPLAIN analysis, health checks, and index recommendations across multiple databases.
README
postgres-mcp
Multi-DB Postgres MCP Server für Cursor / andere MCP-Clients.
Unterstützt Schema-Exploration, sicheres SQL, EXPLAIN, Health-Checks und Index-Empfehlungen.
Voraussetzungen
- Node.js 22+
- Postgres-Zugang (Connection-URI)
- Optional: Docker
Empfohlene Postgres-Extensions (für Slow-Queries / Index-Tuning):
CREATE EXTENSION IF NOT EXISTS pg_stat_statements;
CREATE EXTENSION IF NOT EXISTS hypopg;
pg_stat_statements muss zusätzlich in shared_preload_libraries aktiv sein und Postgres neu gestartet werden.
Schnellstart (lokal)
npm install
npm run build
Eine DB:
DATABASE_URL='postgresql://user:pass@host:5432/dbname' \
ACCESS_MODE=restricted \
npm start
Mehrere DBs:
DB_STAGING_URL='postgresql://user:pass@host:5432/staging' \
DB_PROD_URL='postgresql://user:pass@host:5432/prod' \
DEFAULT_DB=staging \
ACCESS_MODE=restricted \
npm start
HTTP-Transport (für Docker / Remote):
npm run start:http
# -> http://0.0.0.0:8000/mcp
# Health: http://0.0.0.0:8000/health
Konfiguration
Datenbanken
| Variable | Beschreibung |
|---|---|
DATABASE_URL |
Einzelne DB (Alias default) |
DB_<NAME>_URL |
Multi-DB, z. B. DB_STAGING_URL, DB_PROD_URL |
DATABASES |
Alternativ: staging|url,prod|url oder JSON {"staging":"url"} |
DEFAULT_DB |
Default-Alias, wenn mehrere DBs konfiguriert sind |
Tools akzeptieren optional db (Alias aus list_databases). Ohne Angabe wird DEFAULT_DB verwendet.
Access Mode
| Wert | Bedeutung |
|---|---|
unrestricted |
Read/Write (Default, für Dev) |
restricted |
Read-only, 30s Statement-Timeout, kein COMMIT/ROLLBACK |
Setzen über Env oder CLI:
ACCESS_MODE=restricted
# oder
node dist/index.js --access-mode=restricted
Transport
| Variable / Flag | Default | Beschreibung |
|---|---|---|
TRANSPORT / --transport= |
stdio |
stdio oder http |
BIND_HOST / --host= |
0.0.0.0 |
HTTP Bind-Adresse (nicht HOST — Coolify/PaaS überschreibt das) |
PORT / --port= |
8000 |
HTTP Port |
Cursor einbinden
Stdio (lokal)
In Cursor MCP Settings / mcp.json:
{
"mcpServers": {
"postgres": {
"command": "node",
"args": [
"/ABSOLUTER/PFAD/postgres-mcp/dist/index.js",
"--access-mode=restricted"
],
"env": {
"DB_STAGING_URL": "postgresql://user:pass@host:5432/staging",
"DB_PROD_URL": "postgresql://user:pass@host:5432/prod",
"DEFAULT_DB": "staging"
}
}
}
}
Vorher einmal npm run build ausführen.
HTTP (Docker / Remote)
{
"mcpServers": {
"postgres": {
"url": "http://localhost:8000/mcp"
}
}
}
Docker
Build:
docker build -t postgres-mcp .
Run:
docker run --rm -p 8000:8000 \
-e ACCESS_MODE=restricted \
-e DEFAULT_DB=staging \
-e DB_STAGING_URL='postgresql://user:pass@host:5432/staging' \
-e DB_PROD_URL='postgresql://user:pass@host:5432/prod' \
postgres-mcp
Endpoints:
- MCP:
http://localhost:8000/mcp - Health:
http://localhost:8000/health
Hinweis: Wenn Postgres auf dem Host läuft, localhost in der URI je nach OS durch host.docker.internal ersetzen.
Tools
| Tool | Zweck |
|---|---|
list_databases |
Konfigurierte DB-Aliase anzeigen |
list_schemas |
Schemas listen |
list_objects |
Tables / Views / Sequences / Extensions |
get_object_details |
Spalten + Indexes eines Objekts |
execute_sql |
SQL ausführen (max. 50 Rows default) |
explain_query |
EXPLAIN / EXPLAIN ANALYZE / hypopg |
get_top_queries |
Slow Queries (pg_stat_statements) |
analyze_workload_indexes |
Index-Empfehlungen aus Workload |
analyze_query_indexes |
Index-Empfehlungen für konkrete Queries |
analyze_db_health |
Health: index, connection, vacuum, sequence, replication, buffer, constraint, all |
Typischer KI-Flow:
list_databaseslist_schemas/list_objectsget_object_detailsexecute_sqlmitdb: "staging"
Tests
Read-only Integrationstests gegen eine echte DB:
export TEST_DATABASE_URL='postgresql://user:pass@host:5432/dbname'
npm test
Die Tests laufen im restricted-Mode und prüfen alle Tools inkl. Multi-DB und Error-Cases.
Vorlage: .env.test.example
Entwicklung
npm install
npm run build
npm start
npm run start:http
npm test
Build-Output liegt in dist/.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.