OpenAI Codex CLI is an AI developer terminal tool for creating full-stack production-ready applications. Dokploy is an open-source platform-as-a-service that automates infrastructure provisioning and Docker orchestration for self-hosted environments.

This guide explains how to install the Codex CLI, create a full-stack application, and deploy it on Vultr using Dokploy.

Prerequisites

• Have an active OpenAI account, or an API key from the OpenAI developer platform.
• Install Node.js on your local machine.
• Have access to an Ubuntu 24.04 server as a non-root user with sudo privileges.
• Install Dokploy and create your administrator account.
• Set up a domain name, such as app.example.com, and point it to your server's public IP address using an A record.

Set Up Codex CLI

Install the Codex CLI and authenticate with your OpenAI account to prepare the development environment.

1. Install the Codex CLI globally using npm.

   console Copy

   $ npm install -g @openai/codex
   

   Explain Code

2. Authenticate the CLI. If you have a ChatGPT subscription, run the following command and complete the browser login flow.

   console Copy

   $ codex login
   

   Explain Code

3. If you use an API key instead, set it as an environment variable. Replace YOUR_API_KEY with the key from your OpenAI dashboard.

   console Copy

   $ export OPENAI_API_KEY=YOUR_API_KEY
   

   Explain Code

4. Create a new project directory.

   console Copy

   $ mkdir codex-app
   

   Explain Code

5. Navigate into the directory.

   console Copy

   $ cd codex-app
   

   Explain Code

6. Launch the Codex CLI.

   console Copy

   $ codex
   

   Explain Code

Create an Application Using AI

Prompt the Codex agent to generate the project files and container configuration.

1. In the prompt field, describe your application and instruct the agent to make it compatible with Dokploy. Include all required features and the desired service stack. Request a docker-compose.yml if the application involves multiple services.

2. Wait for the Codex agent to generate the application files.

3. Optionally, test the application on your local machine using the generated Docker configuration. The generated README.md file contains local setup instructions.

4. Press Ctrl + C to exit Codex.

5. Provision a new, empty repository on GitHub.

6. Initialize the local Git repository.

   console Copy

   $ git init
   

   Explain Code

7. Stage the generated files for commit.

   console Copy

   $ git add .
   

   Explain Code

8. Commit the files to the local repository.

   console Copy

   $ git commit -m "Initial commit"
   

   Explain Code

9. Set the default branch to main.

   console Copy

   $ git branch -M main
   

   Explain Code

10. Add your GitHub repository as a remote. Replace YOUR_REPOSITORY_URL with your repository's URL.

    console Copy

    $ git remote add origin YOUR_REPOSITORY_URL
    

    Explain Code

11. Push the source code to GitHub.

    console Copy

    $ git push -u origin main
    

    Explain Code

Connect GitHub Repository

Connect your GitHub account with Dokploy to grant it access to pull code from your repositories.

1. Navigate to your instance's IP address on port 3000 (like 192.0.2.1:3000) in a web browser to open the Dokploy dashboard.
2. Log in to the administrator account.
3. Select Git under Settings in the left sidebar.
4. Click GitHub, then click Create GitHub App.
5. Follow the GitHub OAuth prompts to authorize the app and select the repositories Dokploy can access.

Create a New Project

Projects in Dokploy isolate different application environments.

1. Click Projects in the left sidebar.
2. Click Create Project, enter a project name, and click Create.
3. Click on the newly created project to open it.

Connect the Code Repository

Deploy the application by specifying the source repository and creating a new Docker Compose service.

1. On your project page, click Create Service and select Compose.

2. Enter a name for your service and click Create.

3. Open your new Compose service.

4. Locate the Github configuration section.

5. Select your GitHub account and repository.

6. Set the correct Branch.

7. Click Save to apply the repository settings.

8. Open the docker-compose.yml file in your local repository and make the following two adjustments to the web service before deploying:

   • Remove any ports: mappings. Dokploy routes traffic through Traefik, so the web service does not need to bind a port on the host.
   • Add dokploy-network to the service's networks so Traefik can reach the container:
   yaml Copy

   services:
     web:
       networks:
         - default
         - dokploy-network
   
   networks:
     dokploy-network:
       external: true
   

   Explain Code

   Replace web with the actual service name from your compose file.

9. Commit and push the changes to your repository.

Start the Deployment

Dokploy fetches the source code and builds the Docker containers when you trigger a deployment. If your application requires environment variables such as API keys or database credentials, set them under the Environment tab before deploying.

1. Switch to the General settings tab.
2. Click Deploy, then click Confirm.
3. Click View to monitor the real-time deployment logs.

Verify Deployment

Connect your domain for a production-ready URL and encrypted traffic. Dokploy provisions a Let's Encrypt SSL certificate through Traefik when you enable HTTPS.

1. Navigate to the Domains tab and click Add Domain.
2. Select the name of your web service container in the Service Name field, for example web.
3. Type your custom domain into the Host field.
4. Enable the HTTPS toggle.
5. Select Let's Encrypt from the Certificate Provider dropdown.
6. Click Create.
7. Click Validate DNS to verify your domain's DNS configuration.

Add a Database

Dokploy supports deploying managed database containers within the same project. The internal connection URL connects the database directly to your application over Docker's internal network without exposing it to the internet.

1. Navigate to your project page.
2. Click Create Service and select Database.
3. Select the database type that matches your application's requirements.
4. Enter a name for the database and click Create.
5. Open the database service page and click Deploy.
6. Copy the Internal Connection URL.
7. Go to the Environment tab in your application's configuration page.
8. Create or update the DATABASE_URL variable and paste the copied connection URL as the value.
9. Click Save.

Conclusion

You have successfully developed a full-stack application using Codex CLI, pushed the source code to GitHub, and deployed it on Vultr using Dokploy. For more information and advanced configuration options, visit the official Dokploy documentation.