Quickstart

This guide takes you from install to a first real result: a screen designed by Agent, a Figma frame turned into code, or a change to your own project. It takes a few minutes.

Install and sign in

Download Velork from the download page and install it. See Installation for platform details.

The first time you open Velork, you only need to sign in:

1. Click Sign in with browser. Velork opens velork.com in your default browser.
2. Sign in or create an account there. Velork picks up the session automatically.
3. If the browser didn't open, click Open it again next to Didn't see your browser?.

The theme follows your system. You can change it, your keyboard shortcuts, and everything else later in Settings.

Check your AI access

Your account comes with Velork AI credits (2,000 to start on the free Hobby plan), so you can start chatting right away. You can also connect your own AI: a Claude Pro or Max plan through Claude Code, a ChatGPT plan through Codex, a Copilot plan, a Cursor plan with an API key, or an API key from another provider. Credentials stay on your computer.

When your credits run out, Velork shows Connect your AI to keep going. Connect a provider to continue, or wait for paid-plan credits to renew. See AI providers and Plans and usage.

Use the Get started checklist

After sign-in, a Get started card appears under the chat composer. It tracks three steps:

1. Connect your AI: done once you have credits or a connected provider. While you use free credits, this row shows how many are left.
2. Get a first result: any real outcome counts, such as designing a screen, pasting a Figma design, capturing a website, or changing your code
3. Show someone: share a design link, deploy a live link, or commit your work

Open Ways to start on the card for ideas. Steps complete on their own. Close the card with its × button when you're done. To see it again, run Reset Velork Onboarding from the Command Palette.

Open a project

Velork works on folders on your disk. Your repo is the project.

In the chat composer, open the workspace selector and choose Open Folder…, or Connect via SSH… for a remote machine. In the IDE layout you can also use File → Open Folder….

You can skip this step to try the canvas first, but Agent needs a folder to change code.

Get a first result

Pick one starting point.

Design a screen. Describe it in chat:

Design a settings screen for a mobile banking app: profile header,
notification toggles, and a sign-out button.

Agent draws it on the canvas as editable layers.

Start from Figma. Copy a frame in Figma, open the canvas with Cmd + Option + C, click the canvas, and paste with Cmd + V. Then ask Agent to build it:

Implement the selected frame as a React component with Tailwind CSS in this repo.

Change your code. Ask for a concrete change in your project:

Add a dark mode toggle to the header and persist the choice in localStorage.

Agent reads your project before editing. Run your app and open it in the built-in browser with Cmd + Option + 1 to check the result next to the chat.

Review the result

• Canvas: Cmd + Option + C
• Editor: Cmd + Option + E
• Review Changes: Cmd + Option + 4 shows every file Agent changed as a diff

On Windows and Linux, use Ctrl for Cmd and Alt for Option.

Read the diff before you keep it. Ask Agent to run the checks your repo already uses, such as typecheck, lint, or tests.

Plan bigger changes first

For work that spans many files, switch the chat to Plan mode. Press Shift + Tab in the chat input to cycle modes, or pick one from the mode menu. Agent explores the repo, asks questions, and writes a plan before it changes anything. See Modes.

Share it

To share a design, run Copy Design Link from the Command Palette (Cmd + Shift + P). Velork publishes the design if it has no link yet and copies a velork.com/s/ link. Anyone with the link can view, present, and comment.

Next steps