home / mcp / centralmind gateway mcp server

CentralMind Gateway MCP Server

Universal MCP-Server for your Databases optimized for LLMs and AI-Agents.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "centralmind-gateway": {
      "command": "PATH_TO_GATEWAY_BINARY",
      "args": [
        "start",
        "--config",
        "PATH_TO_GATEWAY_YAML_CONFIG",
        "mcp-stdio"
      ],
      "env": {
        "GEMINI_API_KEY": "yourkey"
      }
    }
  }
}

CentralMind Gateway enables you to expose your structured data as fast, secure APIs that AI agents can consume via MCP or OpenAPI 3.1. It analyzes your database schema and samples to automatically generate optimized API configurations, adds security features, and provides easy deployment options for local, on‑premises, or cloud environments.

How to use

Connect your MCP client to the Gateway to start querying your data through REST or MCP. Use the discovery and deployment flow to generate an API configuration based on your database schema, then deploy and start serving requests. You can access generated APIs via the built‑in Swagger UI for documentation and testing, or enable the MCP path to interact with AI tooling that understands MCP metadata and function calling.

How to install

Prerequisites you need before installation are a machine with Docker or a Go toolchain and access to a PostgreSQL or other supported database. Follow these steps to install and run the Gateway.

# Option A: Run with Docker (recommended for quick start)
docker run --platform linux/amd64 -p 9090:9090 \
  ghcr.io/centralmind/gateway:v0.2.18 start \
  --connection-string "postgres://db-user:db-password@db-host/db-name?sslmode=require"

# After starting, you should see the API and MCP endpoints running
# Option B: Build locally (Go)

git clone https://github.com/centralmind/gateway.git
cd gateway

# Install Go dependencies and build
go mod download
go build .

# Start locally with a config file
./gateway start --config gateway.yaml

Additional content

Configuration, security, and usage notes you’ll need to tailor for your environment are included here. Use YAML configuration to customize database connections, tables, endpoints, and MCP metadata. The Gateway supports PII redaction, row‑level security via Lua scripts, and OpenTelemetry‑compatible tracing for observability. You can deploy as a standalone binary, inside Docker, or on Kubernetes, with a YAML configuration guiding behavior and plugin extensions.

Configuration and MCP integration

The Gateway can be configured to expose your data through REST or MCP. To integrate with an MCP client, you run the Gateway as a local process and pass a configuration file that defines your API surface and data access rules. You can also connect via MCP by running a dedicated MCP command that starts the gateway in MCP mode and points to your YAML config.

Example MCP integration (CLI configuration)

{
  "mcpServers": {
    "gateway": {
      "command": "PATH_TO_GATEWAY_BINARY",
      "args": ["start", "--config", "PATH_TO_GATEWAY_YAML_CONFIG", "mcp-stdio"]
    }
  }
}

Available tools

API generation

Automatically creates API configuration based on database schema and sample data using AI providers during the discovery phase.

PII protection

Redacts sensitive data using regex or Presidio plugins to comply with privacy regulations.

MCP protocol support

Implements MCP for seamless integration with AI agents and tool calling.

Observability

Integrates OpenTelemetry to provide request tracing and audit trails.

Caching

Uses time-based and LRU caching to optimize performance for AI workloads.