import type { Metadata } from 'next'import { Geist, Geist_Mono } from 'next/font/google'import { ThemeProvider } from '@/providers/theme'import { QueryProvider } from '@/providers/query'import { Toaster } from '@/components/ui/sonner'import './globals.css' const geist = Geist({ subsets: ['latin'] })const mono = Geist_Mono({ subsets: ['latin'] }) export const metadata: Metadata = { title: 'dFlow', description: 'Ship from your IDE.',} export default function RootLayout({ children }) { return ( <html lang='en'> <body> <ThemeProvider>{children}</ThemeProvider> </body> </html> )}deploy my ecommerce site on AWS EC2 using dflow MCPdflow MCP
dFlow sits between the infrastructure you already pay for and the apps you ship — drive deploys from the dashboard, git, or your AI editor through MCP.
Your cloud
+ bare metal · VPS · on-prem
Your stack
+ any Docker image · 100+ templates
Driven by your AI
Editor deploy
Deploy from the AI editor you already use onto the servers you already pay for.
{ "mcpServers": { "dflow": { "url": "https://app.dflow.sh/api/mcp" } }}Add the dFlow MCP in Cursor, Claude, Codex, or any MCP-compatible editor. One config and your agent gets the full dFlow toolset in chat: apps, services, envs, and deploys.
convert my blog website into a dFlow template by attaching a MongoDB database using dflow mcp
Ask the agent to convert the working tree into a dFlow template, dFlow infers your stack, services, and dependencies and outputs a reproducible template no hand-written YAML.
Tell the agent where to run: Hetzner, EC2, a managed worker, and more. dFlow provisions, builds, and rolls out; you get live deploy logs in chat the moment it goes live.
Agent toolkit
No shell scripts to memorize. dFlow MCP exposes the same operations you would run in the dashboard as tools your editor can call.
Spin up apps, wire git repos, and map domains without leaving the thread.
Add, update, and sync secrets across preview and production environments.
Provision Postgres, MySQL, MongoDB, or Redis beside your app on your worker node.
Kick off builds, watch rollout status, and get preview URLs back in chat.
Pull deployment and runtime logs into the editor when something breaks in prod.
Ask the agent to convert a working repo into a reusable dFlow template.
The new DevOps loop
When a deployed service crashes, stream deployment logs straight into your editor, resolve with your favorite AI tools, apply the fix, and push back to production without leaving your workflow.
Deployment timeline
FAQ
Common questions about MCP setup, supported editors, and shipping from chat.