Skip to Content
Newnadoon AI FrameworkCLI

CLI

Use the Newnadoon CLI to scaffold projects, start the production API server, and optionally run the debug dashboard.

Create a project

Terminal
$ npm create newnadoon@latest

This scaffolds a project and installs @newnadoon/ai-framework (runtime only). The debug dashboard is a separate optional prompt.

Start production API server

Terminal
$ npx newnadoon-api --port 8080

This is the main command. It does not require the dashboard package.

Common flags:

npx newnadoon-api --port 8080 --base-path /api --project-dir ./

Start debug dashboard (optional)

Install the debugger UI first if you skipped it during scaffold:

Terminal
$ npm install -D @newnadoon/ai-framework-dashboard
Terminal
$ npx newnadoon dev

Default dashboard URL:

Dashboard URL
 http://localhost:2024

Voice agents (NewnaVoiceKit)

One-time setup from your project root:

Terminal
$ npx newnadoon-voice setup
Terminal
$ npx newnadoon-voice download-files

Start the voice worker (keep it running while you test):

Terminal
$ npx newnadoon-voice dev

Run the worker in a separate terminal from npx newnadoon dev. See Voice agents for env vars and the sample Voice Assistant walkthrough.

MCP quick client

Terminal
$ npx newnadoon-mcp --baseUrl http://localhost:8080 --tool general-department-agent --input {"task":"hello"}

Use this to quickly test MCP tool calls against your running API server.

Deploy to Newnadoon cloud

With a Portal org API key (write:deploy) and Docker running:

Terminal
$ npx newnadoon deploy --name my-agent --tag latest

In-place update:

Terminal
$ npx newnadoon deploy --name my-agent --tag latest --rollout

Full walkthrough: Host on Newnadoon cloud.

Update framework package

Terminal
$ npm install @newnadoon/ai-framework@latest
APIExamples