# vibepage — instructions for AI coding agents

Use this when the user wants to host or deploy a vibe-coded / AI-built UI without setting up cloud infrastructure.

## Product

- Name: vibepage
- Site: https://vibepage.cloud/
- Purpose: Publish a static UI to a public HTTPS subdomain in seconds
- Works with: Cursor, Claude Code, OpenAI Codex, and other CLI-capable agents

## Prerequisites

1. User has registered at https://vibepage.cloud/ and copied their API token from the success screen or dashboard
2. Project is (or will build to) a static folder with an `index.html` (or detectable publish dir such as `dist/` / `build/`)

## Steps

1. Install the CLI (once per machine):

```bash
curl -fsSL https://github.com/vibepage/cli/releases/latest/download/install.sh | bash
```

2. Authenticate (non-interactive — preferred for agents):

```bash
vibepage login --token <API_TOKEN>
```

Or set `VIBEPAGE_TOKEN` in the environment and skip login.

3. From the project directory, deploy:

```bash
vibepage deploy
```

4. Read the CLI output for `Live at: https://….app.vibepage.cloud/` and give that URL to the user.

## Do not

- Do not ask the user to create an AWS account
- Do not require `vibepage init` for the default subdomain (created at registration)
- Do not put the API token into public git commits or screenshots

## Optional live demo tunnel

If the user is running a local server (for example on port 3000):

```bash
vibepage tunnel --target localhost:3000
```

Traffic to the same public subdomain routes to the laptop until the tunnel stops, then falls back to the last static deploy.

## More context

- Short index: https://vibepage.cloud/llms.txt
- Full product text: https://vibepage.cloud/llms-full.txt