Quickstart

First, what are you building? Either way you hand your coding AI a guide and it does the technical work. Only the finish line is different.

Both paths need a free JustDeploy account and a coding AI on your computer (Codex, Claude Code, or Antigravity).

Deploy an app people visit

No code yet? Start here.

These starters are tiny working apps: a single "Hello World" page (or one API endpoint) plus the DEPLOY.md your AI will need later. You don't edit them by hand. You hand the folder to an AI coding tool and tell it what to build, the same way you'd brief a teammate. The deploy comes after, in the steps below.

  1. Unzip the downloaded file. You'll get a normal folder of source code.
  2. Open that folder in an AI coding tool (Codex, Claude Code, Antigravity). The tool reads every file in the folder, so it knows what it's working with.
  3. Describe what you want in plain words, like "Make this a coffee-tracking page where I log each cup." The AI fills in the code. Keep asking until it looks right.
  1. 1

    Create a project

    In the console, go to Projects → + New. Choose Web if your app has pages people can see, or API if it works behind the scenes with no pages.
  2. 2

    Copy the guide

    Open your project and press Copy. That copies a short guide (DEPLOY.md) to your clipboard, and it already starts with the one line that tells your AI what to do, so you don’t have to type anything.
  3. 3

    Paste it into your coding AI

    Open your coding AI and paste. That’s the whole message. Your AI reads the guide and deploys your app for you. Nothing else to write.
  4. 4

    It goes live

    Your app gets its own URL. Open it, share it, done. Later, to ship a change: make it, then paste the guide again and your AI deploys the change.

Set up a job that runs on a schedule

A Scheduler project runs your code on a repeating schedule instead of serving a website: think nightly reports, cleanups, or reminders. There’s no page to visit; it just runs when you tell it to.

No code yet? Start here.

This starter is a tiny working scheduled job: a handler that already runs on a schedule, plus the notes your AI needs to deploy it. You don't edit it by hand. Hand the folder to an AI coding tool and tell it what the job should do. Setting the schedule comes after, in the steps below.

  1. Unzip the downloaded file. You'll get a normal folder of source code.
  2. Open that folder in an AI coding tool (Codex, Claude Code, Antigravity). The tool reads every file in the folder, so it knows what it's working with.
  3. Describe the job in plain words, like "Every morning, email me yesterday's sales total." The AI fills in the code. Keep asking until it does what you want.
  1. 1

    Create a Scheduler project

    In the console, go to Projects → + New and pick Scheduler. A new scheduler runs hourly by default, and you’ll set your own schedule in step 3.
  2. 2

    Copy the guide and paste it into your AI

    Same as an app: open the project, press Copy, and paste into your coding AI. It reads the guide and deploys your code, so you don’t type any commands.
  3. 3

    Set when it runs

    In the console, describe the schedule in plain words, like “every weekday at 9am”, or pick a preset. You can add more than one time (say, a morning and an evening run), and pause or change it whenever you want.
  4. 4

    It runs automatically

    No URL to visit. It just runs on schedule. Each run’s output shows up in Logs, and Run now triggers one immediately so you can check it works.

Skip the copy-paste with MCP

The paths above hand your AI a guide by copy-paste. Connect your AI to JustDeploy through MCP with a one-time sign-in and no keys, and you can skip even that. From then on, tell your AI “deploy this to JustDeploy” and it creates the project, uploads your code, and returns the live URL or sets up the schedule. Updates work the same way.

If your AI ever seems unsure how JustDeploy works, paste this link into the chat: https://www.justdeploy.ai/llms.txt, a guide written for AIs, covering everything from deploying to databases and troubleshooting.