MCPGate
MCPGate aggregates multiple MCP servers into a single unified endpoint, enabling centralized tool management with granular filtering, automatic namespacing, and observability. Features a real-time web dashboard and optional PostgreSQL-backed audit trails for monitoring and controlling AI tool access across local and remote deployments.
README
MCPGate
Lightweight MCP gateway — aggregate, filter, and observe your MCP tools.
MCPGate sits between your MCP client (Claude Desktop, Claude Code, Cursor, VS Code) and multiple MCP servers. It presents a single unified MCP endpoint while giving you granular control over which tools are exposed.
Features
- Tool Aggregation — Connect multiple MCP servers, expose all tools through one endpoint
- Tool Filtering — Allow/block specific tools per server via simple YAML config
- Tool Prefixing — Automatic namespacing (
github.create_issue) to avoid collisions - Web Dashboard — Real-time status page showing servers, tools, and request logs
- Dual Transport — Stdio (Claude Desktop) or HTTP/SSE (remote clients)
- Audit Trail — Every tool call logged with timing; optional PostgreSQL persistence
- Self-Hosted — Deploy on Railway, Docker, or run locally
Quick Start
Local (stdio — for Claude Desktop)
npx mcp-gate start --config mcpgate.yaml
Add to your claude_desktop_config.json:
{
"mcpServers": {
"mcpgate": {
"command": "npx",
"args": ["-y", "mcp-gate", "start", "--config", "/path/to/mcpgate.yaml"]
}
}
}
Local (HTTP — with dashboard)
npx mcp-gate start --config mcpgate.yaml
# Dashboard at http://localhost:3000
# MCP endpoint at http://localhost:3000/mcp
Docker
docker compose up
# Dashboard at http://localhost:3000
# Includes PostgreSQL for persistent audit trail
Or standalone:
docker run -p 3000:3000 \
-e MCPGATE_CONFIG=$(cat mcpgate.yaml | base64 -w 0) \
-e GITHUB_TOKEN=$GITHUB_TOKEN \
ghcr.io/mprezz/mcpgate
Railway
Add a PostgreSQL plugin for persistent audit trail — MCPGate detects DATABASE_URL automatically.
Configuration
Create a mcpgate.yaml:
gateway:
name: "my-gateway"
transport: "stdio" # stdio | http | both
port: 3000
toolPrefix: true # prefix tools with server name
servers:
- name: "github"
transport: "stdio"
command: "npx"
args: ["-y", "@modelcontextprotocol/server-github"]
env:
GITHUB_TOKEN: "${GITHUB_TOKEN}"
tools:
allow:
- "create_issue"
- "search_repos"
- name: "filesystem"
transport: "stdio"
command: "npx"
args: ["-y", "@modelcontextprotocol/server-filesystem", "/home/user"]
tools:
block:
- "write_file"
- "delete_file"
logging:
level: "info"
See mcpgate.example.yaml for the full reference.
Tool Filtering
Each server supports allow or block (mutually exclusive):
allow— Only these tools are exposed (whitelist)block— All tools EXCEPT these are exposed (blacklist)- Neither — All tools are exposed
Authentication
Add a bearer token to protect the HTTP transport:
gateway:
auth:
token: "${MCPGATE_AUTH_TOKEN}"
When auth is configured:
- All endpoints require
Authorization: Bearer <token>header /healthremains public (for Railway/Docker health checks)- MCP clients pass the token via custom headers in their transport config
Without auth configured, all endpoints are open (for local/private use).
Environment Variable Interpolation
Use ${VAR} or ${VAR:-default} in YAML to reference environment variables:
env:
GITHUB_TOKEN: "${GITHUB_TOKEN}"
API_URL: "${API_URL:-https://api.example.com}"
Dashboard
When running in HTTP mode, MCPGate serves a web dashboard at the root URL:
/— Status page with upstream servers, tools, and request logs/api/status— JSON API for programmatic access/health— Health check endpoint (for Railway/Docker)/mcp— MCP protocol endpoint (Streamable HTTP)
Storage
MCPGate logs every tool call with timing and error information.
- Default — In-memory (no setup required, lost on restart)
- PostgreSQL — Set
DATABASE_URLenvironment variable (tables created automatically)
# Local development with Docker Compose
docker compose up postgres -d
export DATABASE_URL=postgresql://mcpgate:mcpgate@localhost:5432/mcpgate
npx mcp-gate start --config mcpgate.yaml
Environment Variables
| Variable | Required | Default | Description |
|---|---|---|---|
PORT |
No | 3000 |
HTTP port (Railway sets this automatically) |
MCPGATE_CONFIG |
No | — | Base64-encoded YAML config (for Railway/Docker) |
DATABASE_URL |
No | — | PostgreSQL connection string (enables persistent audit trail) |
LOG_LEVEL |
No | info |
debug / info / warn / error |
MCPGATE_AUTH_TOKEN |
No | — | Bearer token for HTTP auth (reference in YAML via ${MCPGATE_AUTH_TOKEN}) |
Development
git clone https://github.com/mprezz/mcpgate.git
cd mcpgate
npm install
npm run dev # development with hot reload
npm run build # compile TypeScript
npm run test # run tests (vitest)
npm run lint # eslint
npm run typecheck # tsc --noEmit
Architecture
Client (Claude, Cursor)
│
▼ stdio or HTTP/SSE
┌──────────────────────────────┐
│ MCPGate │
│ │
│ ┌─────────────────────┐ │
│ │ Tool Registry │ │ ← filters + namespaces tools
│ └─────────────────────┘ │
│ ┌─────────────────────┐ │
│ │ Tool Router │ │ ← routes calls to correct server
│ └─────────────────────┘ │
│ ┌─────────────────────┐ │
│ │ Upstream Manager │ │ ← manages server connections
│ └─────────────────────┘ │
│ ┌─────────────────────┐ │
│ │ Storage │ │ ← memory or PostgreSQL
│ └─────────────────────┘ │
└──────────────────────────────┘
│ │ │
▼ ▼ ▼
Server A Server B Server C
(GitHub) (Filesystem) (Custom)
License
Apache 2.0 — see LICENSE
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.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
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.