Join us for Render's 6/18 conference in SF

Get your ticket
Get Started
Back to agents

Claude Code + Render

Deploy and manage your Render apps from Claude Code.

Claude Code is Anthropic's terminal-based coding agent. Paired with Render, it can deploy web services, create Render Postgres databases, inspect logs and metrics, update environment variables, and help plan infrastructure changes from your shell. The combination is well-suited to long-running production apps: Render's managed infrastructure, real Postgres with HA and read replicas, persistent disks, background workers, no cold starts, and render.yaml Blueprints for infrastructure-as-code mean Claude has stable surfaces to act against.

Agent details
Agent
Claude Code
Short name
Claude
Vendor
Anthropic
Install
npm install -g @anthropic-ai/claude-code
Start
claude
Skills mode
Agent-invoked
Docs
Open Claude Code docs

Quick start

  1. Install Claude Code globally.

  2. Add the Render MCP server to Claude Code. Generate an API key in your account settings, then run:

  3. Install Render's official skills so Claude knows how to deploy, debug, and monitor:

    The installer auto-detects Claude Code and installs the right files in ~/.claude/skills/. If you already have the Render CLI installed, render skills install works too. See the Render skills section below for both options.

  4. Start Claude Code.

  5. Set your Render workspace. If your Render account has multiple workspaces, tell Claude which one to operate against. For example, "Set my Render workspace to <YOUR_WORKSPACE_NAME>."

  6. Deploy a template. Try a prompt like "Deploy the Next.js + Postgres template from render.com/templates and walk me through any choices."

Quick links

Skills on GitHub MCP Server on GitHub The Render CLI docs Claude Code docs

MCP configuration

The claude mcp add command from Quick start writes an entry into ~/.claude.json.

MCP setup
Install command
claude mcp add --transport http render https://mcp.render.com/mcp --header "Authorization: Bearer $RENDER_API_KEY"
Config file
~/.claude.json
OAuth
Not supported yet

The equivalent JSON looks like this:

The Render MCP server currently uses API-key authentication. Generate a key in your account settings and pass it as a Bearer token. OAuth support is on the roadmap.

For full setup details, including troubleshooting, scoping a key to a single workspace, and the complete tool reference, see the Render MCP server repo and the MCP server docs page.

The prompts below can use a combination of MCP tools, Render skills, the Render CLI, and guided Dashboard or API steps.

Example prompts

Start with one of these prompts, then refine the plan with your agent.

Deploy

Deploy this project to Render.

Data

Create a Render Postgres database and wire DATABASE_URL to my web service.

Debug

Check the deploy logs for the last failed deploy and fix the error.

Scale

Review my service's scaling settings and recommend the safest path to autoscaling at 70% CPU.

Domain

Set up a custom domain with SSL for app.example.com.

Blueprint

Create separate staging and production environments using a render.yaml Blueprint.

AI app

Deploy an AI chatbot with Postgres for chat history and a background worker for embeddings.

Template

Set up n8n with Postgres and a persistent disk on Render.

SaaS

Deploy a Next.js SaaS starter with Auth.js and Stripe webhooks.

Workflows

Set up a Workflow to orchestrate my data pipeline with retries and subtasks.

Render skills

Render's official skills give Claude Code the procedural know-how to deploy services, debug failed builds, run read-only database queries, and manage infrastructure changes. Skills are agent-invoked SKILL.md files that Claude Code picks automatically based on your prompt. They can combine MCP tools with the Render CLI, local code edits, and Dashboard or API guidance for workflows that MCP alone doesn't cover.

Install skills

Works without the Render CLI installed. The installer detects supported tools on your system (Claude Code, Codex, Cursor, OpenCode) and prompts you to select which skills to install.

Requires version 2.10 or later of the Render CLI.

Check your current version with render --version.

Use this if you already have the Render CLI on your system. It's the same flow as the npx installer, just routed through the CLI.

Skills catalog

View Render skills on GitHub

Common workflows

Deploy a Postgres database

  • Ask Claude Code to create a Render Postgres database and pick a plan.
  • Store the database's internal connection string as a service env var (DATABASE_URL).
  • Run any required migrations from your service's pre-deploy command.
  • For read-heavy workloads, ask it to evaluate whether high availability or a read replica belongs in your plan, then make that change from the Render Dashboard or API.

Deploy a project from a Blueprint

  • Ask Claude Code to read the project and propose a render.yaml Blueprint.
  • Review the proposed services, databases, and env vars.
  • Validate the Blueprint with the Render CLI, then apply it from the Render Dashboard or API.
  • Confirm the first deploy succeeds by checking logs and the health-check endpoint.

Debug a failing deployment

  • Ask Claude Code to fetch the most recent failed deploy's logs.
  • Have it identify the root cause (missing env var, port binding, OOM, build error).
  • Apply the fix by updating env vars, editing code, or changing service settings through the Render Dashboard or API.
  • Push the code fix or update the service settings, then watch the next deploy until the service is healthy.

Configure environment variables and secrets

  • Ask Claude Code to list the current env vars on the target service.
  • Add or update values, marking sensitive secrets with sync: false.
  • For values shared across services, ask it to propose an env group change.
  • Confirm a new deploy starts, or redeploy from the Render Dashboard if needed.

Set up a custom domain

  • Ask Claude Code to walk through adding a custom domain to the service.
  • It returns the required DNS record (CNAME or A/ALIAS for an apex).
  • Add the DNS record at your registrar.
  • Confirm verification and TLS issuance with a final status check.

Tips for success

  • Install the Render MCP server so Claude Code can read service state, fetch logs and metrics, and apply supported env-var changes instead of guessing from your local repo.
  • Define your infra in a render.yaml Blueprint. Agents read and modify Blueprints reliably, and the file becomes the source of truth that your team and your agent share.
  • Ask for service status before deploying. A quick "what's the current state of my services?" prompt catches mismatches between local intent and what's actually deployed before you ship a change.

FAQ

Troubleshooting

claude mcp command fails or hangs

Confirm Claude Code is at least version 1.0.30 with claude --version. The HTTP MCP transport used here was added in that release. Upgrade with npm install -g @anthropic-ai/claude-code@latest.

Authentication errors from the Render MCP

The Authorization header must be the literal string Bearer <KEY>. If your shell didn't expand $RENDER_API_KEY, the header will contain the literal $RENDER_API_KEY and the server will reject it. Run echo $RENDER_API_KEY to confirm the variable is set, or paste the key value directly into the header.

Claude can't find the render CLI

The render skills install command needs the Render CLI on your PATH. Install it with Homebrew (brew install render) or the install script, then confirm with render --version.

Community

Start a new project Read the docs

Related agents

Other coding agents that pair well with Render.