Deploying your app

gy deploy is the only command you need to ship. It auto-detects your stack, generates a Dockerfile when one is missing, builds, uploads, and deploys.

The basics

# Deploy from the current directory
gy deploy

# Deploy a specific directory
gy deploy ./my-app

# Interactive guided setup
gy deploy --wizard

Aliases: gy d, gy push, gy up.

What gy deploy detects

Type	Detector	Example frameworks
Node.js	package.json	Next.js, Express, Vite, Nuxt
Python	requirements.txt, pyproject.toml	Django, Flask, FastAPI, Streamlit
Go	go.mod	Gin, Fiber, Echo, Chi
Rust	Cargo.toml	Any Cargo project
Static	index.html	Plain HTML
Docker	Dockerfile	Your own container

Flags

Flag	Description
-n, --name	App name (defaults to the directory name)
-p, --project	Project name (auto-created if new)
--port	Port your app listens on (auto-detected)
--tier	Resource tier: t1, t2, or t3 (default t1)
--domain	Custom domain (empty = auto-generated)
--public	Make the app publicly accessible (default true)
--dockerfile	Path to a Dockerfile (auto-generated if absent)
-w, --wizard	Interactive guided setup
--no-ai	Disable AI-assisted detection
--ai-key	GenAI API key for AI-assisted detection (or set GY_AI_KEY)
--json	Emit structured JSON output

Examples

# Name the app and pick a tier
gy deploy --name "my-cool-api" --tier t2

# Deploy into a specific project
gy deploy --project team-ops

# Explicit port
gy deploy --port 8080

# Deploy a directory with a script-friendly JSON result
gy deploy ./backend --json

The .gy.json file

After a successful deploy, gy writes a .gy.json file to your project:

{
  "app": "my-cool-api",
  "project": "default",
  "port": 3000,
  "publicAccess": true,
  "tier": "t1",
  "env": {
    "DATABASE_URL": "postgres://…"
  }
}

It contains no UUIDs — everything uses friendly names, and the CLI resolves identifiers internally. Commit it so teammates and CI can use the same app.

Re-deploying

Run gy deploy again any time. gy reads .gy.json, targets the same app, and ships your changes.