Render.com에 직접 AI 보조를 통해 배포합니다.

이 MCP (Model Context Protocol) 서버는 Claude와 같은 AI 조수를 사용하여 Render API와 상호 작용하고 Render.com에 서비스의 배포 및 관리 할 수 있습니다.

특징

이 서버는 모든 렌더링의 공식 MCP 서버를 다룹니다 ** mutating 서버 의도적으로 omits (triggering deploys, deleting resource, 사용자 정의 도메인 관리, 재시작 및 취소 배포).

서비스

• 모든 서비스를 나열하고 특정 서비스의 세부 정보를 얻을
• 서비스 만들기: 일반 create_service 플러스 타입 create_web_service, create_static_site 및 create_cron_job]
• 배포, 재시작 및 삭제 서비스
• 환경 변수 및 사용자 정의 도메인 관리

**직업 **

• 배포 기록을 얻고 단일 배포를 얻고, in-progress 배포를 취소

** 워크스페이스 **

• Workspaces 목록, workspace 세부 정보를 얻고, default workspace를 선택하십시오. create/logs/metrics 도구는 ownerId]를 호출할 필요가 없습니다.

**보관 **

• 목록 및 필터 로그 (list_logs) 및 로그 라벨 값을 enumerate
• Fetch 성능 메트릭 ([T6]]): CPU, 메모리, HTTP 요청/경량,

Render MCP Server

Deploy to Render.com directly through AI assistants.

This MCP (Model Context Protocol) server allows AI assistants like Claude to interact with the Render API, enabling deployment and management of services on Render.com.

Features

This server covers everything Render’s official MCP server does plus mutating operations their server intentionally omits (triggering deploys, deleting resources, managing custom domains, restarting, and cancelling deploys).

Services

• List all services and get details of a specific service
• Create services: generic create_service plus typed create_web_service, create_static_site, and create_cron_job
• Deploy, restart, and delete services
• Manage environment variables and custom domains

Deploys

• Get deployment history, get a single deploy, and cancel an in-progress deploy

Workspaces

• List workspaces, get workspace details, and select a default workspace so create/logs/metrics tools don’t need an ownerId each call

Observability

• List and filter logs (list_logs) and enumerate log label values
• Fetch performance metrics (get_metrics): CPU, memory, HTTP requests/latency, bandwidth, instance count, active connections

Datastores

• Postgres: list, get, create, and run read-only SQL queries (query_render_postgres)
• Key Value (Redis): list, get, and create instances

Note on query_render_postgres: this tool connects to your database using the pg driver and enforces read-only access (single statement, SELECT/WITH/ EXPLAIN/SHOW only, run inside a READ ONLY transaction). Run npm install after installing the package to ensure the pg dependency is present.

Installation

npm install -g @niyogi/render-mcp

Configuration

1. Get your Render API key from Render Dashboard
2. Configure the MCP server with your key:

node bin/render-mcp.js configure --api-key=YOUR_API_KEY

Alternatively, you can run node bin/render-mcp.js configure without the --api-key flag to be prompted for your API key.

Usage

Starting the Server

node bin/render-mcp.js start

Checking Configuration

node bin/render-mcp.js config

Running Diagnostics

node bin/render-mcp.js doctor

Note: If you’ve installed the package globally, you can also use the shorter commands:

render-mcp start
render-mcp config
render-mcp doctor

Using with Different AI Assistants

Using with Claude Code

The quickest way is the claude mcp add CLI, which registers the server via npx (no global install needed):

claude mcp add render -e RENDER_API_KEY=YOUR_API_KEY -- npx -y @niyogi/render-mcp start

By default this adds the server to the current project. Use a scope flag to change that:

• -s user — available across all your projects