> ## Documentation Index
> Fetch the complete documentation index at: https://docs.origami.techmitten.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Get Started with Origami AI: From Zero to First Video

> Open the hosted app or self-host with npm. This guide walks you through every path to your first AI-generated video in Origami AI — no sign-up needed.

This guide shows you the fastest path from zero to your first project in Origami AI. You can start immediately using the hosted app — no sign-up, no installation — or run the full studio locally on your own machine with a few terminal commands. Either way, you'll be creating AI-narrated videos, short-form content, and more within minutes.

## Steps to Get Started

<Steps>
  <Step title="Open the hosted app">
    The quickest way to start is the live hosted version — no account or installation required:

    👉 **[https://origami.techmitten.com/](https://origami.techmitten.com/)**

    Open that URL in Chrome or Edge 113+ and you're ready to go. Skip ahead to step 3.
  </Step>

  <Step title="(Optional) Self-host locally">
    Prefer to run Origami AI on your own machine? Make sure you have **Node.js ≥ 20.19.0** and a WebGPU-capable browser installed, then run:

    ```bash theme={null}
    git clone https://github.com/TechMitten/Origami-AI.git
    cd Origami-AI
    npm install
    npm run dev
    ```

    Open **[http://localhost:3006](http://localhost:3006)** in your browser once the server starts.
  </Step>

  <Step title="Wait for the first-run model download">
    On your first visit, Origami AI automatically downloads the AI models it needs (the default text model and the Kokoro.js TTS engine). This may take a few minutes depending on your connection speed.

    Once downloaded, the models are **cached in your browser's storage** — subsequent launches are instant, even offline. You can monitor download progress in the status bar at the bottom of the app.
  </Step>

  <Step title="Choose a tool and build something">
    Use the navigation sidebar to pick your first project type:

    * [**PDF to Video**](/features/pdf-to-video) — upload a PDF deck and generate a narrated video
    * [**Shorts Generator**](/features/shorts-generator) — type a topic and produce a vertical short
    * [**Screen Recording**](/features/screen-recording) — capture your screen with auto-zoom
    * [**AI Assistant**](/features/ai-assistant) — chat with a local AI model in your browser
  </Step>
</Steps>

## Available Scripts

If you're running Origami AI locally, these npm commands cover your full development and deployment workflow:

| Command           | Description                                                                                 |
| ----------------- | ------------------------------------------------------------------------------------------- |
| `npm run dev`     | Starts the Express + Vite dev server with hot module replacement at `http://localhost:3006` |
| `npm run build`   | Builds the production client bundle into the `dist/` directory                              |
| `npm run preview` | Serves the production build locally so you can test before deploying                        |
| `npm run lint`    | Runs ESLint across the codebase to catch style and type issues                              |
| `npm run stop`    | Terminates any active process running on port 3006                                          |

## Run with Docker

If you'd rather use Docker for a self-contained environment with better resource isolation — especially useful for heavy video rendering — a single command gets you running:

```bash theme={null}
docker compose up --build
```

The app will be available at **[http://localhost:3000](http://localhost:3000)** once the container starts.

<Tip>
  For the best WebGPU experience and full access to every Origami AI feature, use **Chrome or Edge 113+**. These browsers have the most complete and stable WebGPU implementation. Firefox Nightly (with the `dom.webgpu.enabled` flag) and Safari 18+ on macOS Sonoma are also supported but may have minor limitations with certain GPU-accelerated effects.
</Tip>