Getting Set Up with Insites v6 and AI

This guide takes you from nothing to a live page on your own Insites instance, built and deployed by an AI assistant. You describe what you want, the AI writes the code, and Insites CloudShell deploys it. There are two paths below: a quick start for developers who already use an AI coding tool, and a step-by-step walkthrough for everyone else.

What You Need

• An Insites account - sign up at console.insites.io.
• An Insites instance - one you created, or one you have been given access to with an Edit role or above. CloudShell only shows instances you can change.
• An AI assistant that supports remote MCP servers - for example Claude Code, Claude, Cursor, Windsurf, or ChatGPT. This guide uses Claude Code as the worked example.
• A paid AI plan - Claude Code and custom connectors in Claude need a paid Claude plan. Other AI assistants have their own plan requirements.

Quick Start for Developers

If you already use an AI coding tool, these five steps are all you need.

1. Sign up at console.insites.io and create an instance.
2. Add CloudShell to your AI assistant as a remote MCP server: https://cloudshell.insites.io/mcp. In Claude Code:

   claude mcp add --transport http insites-cloudshell https://cloudshell.insites.io/mcp

3. Sign in when prompted. In Claude Code, run /mcp, choose insites-cloudshell, and authenticate in the browser window that opens.
4. Give your AI the Insites Logic Engine so it builds the Insites way.
5. Ask it to list your instances, select one, and deploy. To keep every change in version control, connect a GitHub repository.

Step-by-Step Walkthrough

New to Insites or to AI coding tools? Follow these steps in order. Each one takes a few minutes.

Step 1: Create Your Account and Instance

Go to console.insites.io and sign up for an account. Once you are in, create a new instance. This is where your website or application will live, at an address like https://your-instance.insites.site.

Step 2: Install an AI Assistant

Install the AI assistant you want to use. We recommend Claude Code, which runs in your terminal or inside VS Code and can read and write the files in your project folder. Follow the install instructions on its website, then create an empty folder for your project and open Claude Code in it.

Any AI assistant that can connect to a remote MCP server will work. MCP (Model Context Protocol) is the open standard that lets an AI assistant use outside tools, such as CloudShell.

Step 3: Connect CloudShell

CloudShell is how your AI assistant talks to your instance. You connect it once, and it works in every conversation after that.

In Claude Code, run this command in your terminal:

claude mcp add --transport http insites-cloudshell https://cloudshell.insites.io/mcp

Then start Claude Code, type /mcp, choose insites-cloudshell, and select Authenticate. A browser window opens. Sign in with the same Insites account you use for the Console.

In other AI assistants, look for the setting to add a remote MCP server or a custom connector, and enter this address:

https://cloudshell.insites.io/mcp

Sign in with your Insites account when the assistant asks you to. There is nothing else to configure: no passwords, keys, or instance IDs go into any settings file.

Step 4: Choose Your Instance

Ask your AI assistant:

List my Insites instances and select the one called your-instance.

It shows every instance you have an Edit role or above on, and selects the one you name. Every deploy goes to the selected instance until you choose another.

Step 5: Give Your AI the Insites Logic Engine

The Insites Logic Engine is a public set of rules and reference docs that teaches AI assistants how Insites works. Paste this into your AI assistant at the start of your project:

Before you write any code for this project, read the Insites Logic Engine at
https://github.com/insites-io/insites-logic-engine

Start with skills/insites/SKILL.md and follow its decision trees, then use the
reference docs in skills/insites/references/ and the rules in logic-engine/rules/.
Whenever you are unsure how something works on Insites, check the Logic Engine
first instead of guessing, and follow its conventions over your own assumptions.

Step 6: Build and Deploy Your First Page

Now ask for something real. For example:

Build a simple homepage for my instance with a heading, a short welcome
paragraph, and a contact button. Deploy it to my selected instance, then
check the logs for any errors.

Your AI writes the page, deploys it through CloudShell, and confirms when it is done. Open your instance address, for example https://your-instance.insites.site, and you will see your new homepage. From here, keep describing what you want and your AI will keep building.

Working with GitHub

Deploying directly is the fastest way to start. Once a project grows, or more than one person works on it, connect a GitHub repository so every change is a commit with a history you can review and undo.

1. Connect the repository. Ask your AI: Connect GitHub to my instance. It gives you a link to install the Insites GitHub app. Choose the repository to use. Your instance then watches one branch of it for changes.
2. Push changes. Your AI commits files to the repository instead of deploying them directly. Every push to the watched branch deploys to your instance automatically.
3. Work on branches. Ask your AI to push work to a feature branch. It deploys only when merged into the watched branch. You can also change which branch is watched, for example a staging branch for a staging instance.
4. Review and undo. Ask your AI to list recent commits and GitHub deploys. To undo a change, ask it to revert a commit, then deploy the reverted state.

Next Steps

• About CloudShell - what CloudShell is and how it works.
• CloudShell Tools - every tool your AI can use, with example prompts.
• Insites Logic Engine - how your AI learns to build on Insites.
• Development Workflow - environments and the full development cycle.