AppWizzy Launches Projects MCP for AI Coding Agents

We’ve launched AppWizzy Projects MCP, creating a direct connection between AI coding agents and the AppWizzy project infrastructure.

Your agent can now take code from your workspace, create an AppWizzy project, help you select the right machine, and monitor the project until it is running.

One connection, one workflow: local code → AppWizzy project → live application.

What Your Agent Can Do

With AppWizzy Projects MCP, an AI agent can manage the project creation workflow without manual API requests or copied bearer tokens.

It can:

• Show available machines with their specifications and pricing.
• Create a new AppWizzy project from a ZIP archive.
• List your existing AppWizzy projects.
• Check project creation progress.
• Monitor deployment status.
• Return the project URL when the application is ready.

The agent can guide you through the process, explain the available infrastructure options, and ask for confirmation before reserving credits.

How ZIP Import Works

Your ZIP archive is imported into AppWizzy’s external-code template. This allows you to bring an existing application-or code generated locally by an AI agent, into a complete AppWizzy environment without rebuilding the project manually. Temporary upload data is removed after the archive has been processed.

Connect AppWizzy to Codex

Codex Desktop

Open: Settings → MCP servers → Add server

Then enter the following values:

Save the server and click Authenticate. Your browser will open the AppWizzy authorization page, where you can approve access to your account.

OAuth Authorization, Without Bearer Tokens

Your AppWizzy password is never shared with Codex. Authentication happens directly in your browser through OAuth, so there is no need to generate, copy, or expose a bearer token inside your AI conversation or configuration.

Connect Through Codex CLI

Run:

codex mcp add appwizzy_projects \ --url https://appwizzy.com/mcp/projects \ --oauth-client-id appwizzy-codex-cli

Then authenticate:

codex mcp login appwizzy_projects

The login command will open the AppWizzy authorization flow in your browser.

Manual Configuration

You can also configure the MCP server manually in:

~/.codex/config.toml

Add:

[mcp_servers.appwizzy_projects]url = "https://appwizzy.com/mcp/projects"[mcp_servers.appwizzy_projects.oauth]client_id = "appwizzy-codex-cli"

Restart Codex after saving the file.

Create Your First Project

Once the connection is active, give your agent a direct instruction.

Example Prompt

Show me the available AppWizzy machines with their specifications and pricing. Then create a project from ./my-project.zip. Ask me to confirm the machine before reserving credits, and monitor the project until it is ready.

The agent will:

1. Inspect the available machine options.
2. Explain their specifications and cost.
3. Ask you to select or confirm a machine.
4. Upload your application code.
5. Create the AppWizzy project.
6. Monitor project creation and deployment.
7. Return the project link when it is ready.

Why This Matters

AI coding agents are already capable of writing complete applications. But generating code is only part of the job. The application still needs somewhere to run. Infrastructure must be selected, a project must be created, the code must be uploaded, and the deployment must be monitored. Until now, moving from an AI coding conversation to a real running environment still required a series of manual steps.

AppWizzy Projects MCP closes that gap. Your agent writes and prepares the code. AppWizzy provides the runtime, infrastructure, project lifecycle, and deployment environment.

From Local Code to a Running Project

The goal is straightforward: make launching an application a natural continuation of building it with an AI agent.

Connect AppWizzy Projects MCP, point your agent to a ZIP archive, choose a machine, and launch your first agent-created project.