How cervo works

GETTING STARTED

Connect your AI tool — Claude Code, a desktop assistant, anything that speaks MCP — to cervo's server:

Ask it to authenticate with cervo. You will be asked to confirm your email address, and a six-digit code lands in your inbox; paste it back into the chat. The confirmed address owns everything you create, and the chat stays signed in for four hours.

Then ask for a website. A site's name is its slug — lowercase letters, digits, and hyphens — and becomes its address:

Within seconds the site is live at https://my-cool-site.cervo.click, serving a default page until you publish your own.

UPDATING YOUR SITE

Every new site starts with cervo's default page. It is written only when your site has no index.html of its own — the moment you publish your files, they are yours, and no redeploy will ever overwrite them.

Updating is the same conversation as creating:

HOW DEPLOYMENTS WORK

Creating a website queues a deployment, which a background worker runs as a chain of steps the chat can follow in real time: create the site's directory and write the default page if none exists, regenerate the web server's configuration from the database, and reload it.

A deployment's status runs pending, then deploying, then live — or failed, with the error shown when you list your websites. Failures retry on their own, and asking to create your own failed site again queues a fresh deployment. Every step is idempotent, so retrying is always safe.