create

create scaffolds a new application from an official Machina starter — the code lands on your machine, already wired to talk to a Machina pod, with deployment workflows for AWS and Azure included.

This is the local, code-first counterpart to Factory: Factory builds and deploys an app for you in a sandbox and opens a pull request; create hands you the repository and you take it from there.

Usage

machina create ai-app <name>                    # scaffold into ./<slug>
machina create ai-app "My App" -d ./my-app      # choose the output directory
machina create ai-app my-app --no-git           # skip git init
machina create ai-app my-app --json             # machine-readable output

Subcommands

Command	Description
create ai-app <name>	Create a chat app wired to a Machina pod and ready for AWS or Azure

ai-app

Option	Short	Description
--directory	-d	Output directory. Defaults to ./<slug>
--template-ref		Branch or tag of the official boilerplate. Default: main
--no-git		Don't initialize a Git repository
--json	-j	Output machine-readable JSON

The app name is slugified for the directory and for placeholders inside the template (lowercased, non-alphanumerics collapsed to -, truncated to 63 characters). The command refuses to write into a directory that already has files in it.

Source template: machina-sports/machina-boilerplate.

After scaffolding

cd <your-app>
cp .env.example .env.local
npm install && npm run dev

Git is initialized on a main branch unless you passed --no-git (if Git isn't available the command says so and continues — the files are still written). Deployment to AWS or Azure runs through the GitHub Actions workflows in the generated repo; the variables they need are documented in the app's own DEPLOYMENT.md.

Related

• factory — build and ship an app from a prompt, in a sandbox.
• project — the Machina project the generated app talks to.
• credentials — the API key the app authenticates with.