Lovable MCP server

1. 1

   Ensure your Lovable account is on any plan and that 'Third-party MCP client access' is enabled for your workspace.

   This setting allows external AI tools to connect to your Lovable account. It's usually on by default for most plans, but Enterprise users might need to turn it on.

2. 2

   Open your preferred supported AI client (ChatGPT, Claude, Cursor, or VS Code).

   You need one of these specific AI tools because they are currently the only ones that can properly connect to the Lovable MCP server.

3. 3

   Follow the connection instructions specific to your AI client to add the Lovable MCP server.

   Each AI client has a slightly different way to add external tools. This step links your AI client to your Lovable account using a secure sign-in process.

4. 4

   Once connected, start giving your AI agent prompts to create, modify, or deploy Lovable projects.

   Your AI agent can now use Lovable's capabilities directly. For example, you can ask it to 'Create a Lovable project called 'Feedback Hub' with a form for collecting user feedback'.

Create a Lovable project called 'My First AI App' with a simple landing page that has a sign-up form for a newsletter.

Model Context Protocol (MCP)

This is a standard way for AI tools to find out what other software can do and then use those features. Think of it like a universal remote control for AI, letting it talk to different apps.

OAuth flow

This is a secure way for you to give one application (like your AI client) permission to access your information or perform actions in another application (like Lovable) without sharing your password directly.

Tool calls

These are specific commands or actions that your AI agent can ask Lovable to perform, like 'create_project' to make a new project or 'deploy_project' to publish an app.

Unified diff

This is a way to show the differences between two versions of a text file, like code. It highlights what was added, removed, or changed, making it easy to see revisions.

• Connecting an AI client gives it the same access to your Lovable account as you have, meaning it can see, edit, and deploy all your projects and use your credits.
• Using 'deploy_project' will publish your app to a live URL, which might be publicly accessible depending on your plan and settings.
• The 'query_database' tool runs SQL commands with your full database permissions, allowing it to read, write, and change your database structure.