create-pyai-app

Scaffold a runnable PyAI example in one command — no clone, no setup ceremony.

npm create pyai-app@latest                      # interactive: pick an example
npm create pyai-app@latest openai-drop-in        # by name
npx create-pyai-app voice-cloning my-app --key pyai_test_… --install

It pulls the chosen example straight from the public pyai-examples repo, seeds .env from the template (with your key if you pass --key), and prints the next steps.

Usage

npm create pyai-app@latest [example] [dir] [options]

| Argument / option | Meaning | |---|---| | example | Which example (e.g. openai-drop-in). Omit to choose interactively. | | dir | Target directory (default: the example's name). | | --key <k> | Write PYAI_API_KEY=<k> into .env. A pyai_test_… sandbox key is fine. | | --install, -i | Run npm install after scaffolding. | | --force, -f | Scaffold into a non-empty directory. | | --list, -l | List the available examples and exit. |

Examples it can scaffold

The list is read live from the repo, so it's always current. A few highlights:

• openai-drop-in — your OpenAI audio code, served by PyAI (change the base URL).
• voice-cloning — clone a voice end to end and use it in Omni.
• pyai-site-voice-concierge — a browser voice agent (one-click deployable).
• recap-call-intelligence — batch transcribe → talk-ratio + keywords + summary.

Run npm create pyai-app@latest --list for the full set.

Requirements

• Node ≥ 18 (uses built-in fetch). Zero runtime dependencies.
• A PyAI key from the console when you run the example (pyai_test_… works everywhere with sandbox caps).