Skip to main content
Install Base44 skills and let your AI coding agent create, build, and deploy your project.
You need Node.js 20.19.0 or higher installed.

​
Setup

1

Install Base44 skills

Install skills globally so they are available in any project:
npx skills add base44/skills -g
This teaches your AI coding agent how to work with Base44 projects, including the CLI, SDK, entities, backend functions, and deployment.
2

Open your AI coding tool

Open a folder in Cursor, Claude Code, Codex, OpenCode, or any agent that supports skills. The Base44 skills are automatically available.
3

Describe what you want to build

Give your agent a prompt describing your project. Be specific about the data you need, how users interact with it, and any features you want. Here are some examples:“Create a new Base44 project for a job application tracker. I need to store data for Companies, Applications, and Interviews. Applications should have status, position title, salary range, and notes. Add authentication so each user only sees their own data.”“Create a new Base44 project for a customer support tool. I need an AI agent that can look up orders by email, check delivery status, and create refund requests. Store conversations and link them to customer records.”“Create a new Base44 project for a team standup bot. Add a Slack connector and a backend function that runs daily, collects standup updates from a Standups table, and posts a summary to a Slack channel.”
Your agent may prompt you to authenticate with Base44 during setup. This opens a browser window for you to sign in.

​
What your agent will do

When you give your agent a prompt, it will typically:
  1. Install the Base44 CLI if it is not already installed.
  2. Run base44 create to scaffold the project.
  3. Define entity schemas based on your description.
  4. Configure AI agents, OAuth connectors, and other resources.
  5. Write frontend and backend code.
  6. Set up authentication, RLS rules, and other settings.
  7. Deploy the project with base44 deploy.
You can guide your agent at each step or let it work through the full process.

​
Keep building

After the initial setup, your agent can continue to make changes across your project. Some things you can ask it to do:
  • Add or modify entity schemas and push them with the CLI.
  • Write backend functions with automations and cron schedules.
  • Configure AI agents with custom tools and data access.
  • Set up OAuth connectors for external services.
  • Build frontend pages and components using the SDK.
  • Deploy updates to production.

​
Tips for better results

  • Be specific about your data model: Name your entities and describe the fields you need, including relationships between them.
  • Mention authentication early: If your app needs login or per-user data, say so in your initial prompt so the agent sets up RLS from the start.
  • Describe behavior, not just structure: Instead of “add a notifications entity,” try “send an email notification when a task is assigned to someone.”
  • Iterate: Start with a basic version and ask your agent to add features one at a time.

​
See also