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

# Setup & Requirements

<Warning>
  Beeble Studio currently supports **Windows and Rocky Linux only**, and
  requires an **NVIDIA GPU**.
</Warning>

## Installation Guide

### 1. Download & Install

<Card title="Download Beeble Studio Installer" icon="download" href="https://beeble.ai/resources/downloads">
  Get the installer for Windows or Rocky Linux
</Card>

### 2. Launch & Sign In

Launch the app and sign in. You'll need an active Beeble Studio plan (7-day free trial available).

### 3. Download Models

Download required engine and model files. You're now ready for local GPU processing! 🎉

***

## Hardware Requirements

| Component               | Minimum                                                       | Recommended                                                            |
| ----------------------- | ------------------------------------------------------------- | ---------------------------------------------------------------------- |
| **OS**                  | Windows 10/11 ¹ ²<br />Rocky Linux 8 & 9                      | Windows 10/11 ¹<br />Rocky Linux 8 & 9                                 |
| **CPU**                 | Intel or AMD (AVX support, 2020 or newer)                     | Intel or AMD (AVX support, 2020 or newer)                              |
| **NVIDIA GPU**          | Ampere architecture or newer (e.g., RTX 30-series, RTX A6000) | Ada Lovelace architecture or newer (e.g., RTX 50-series, RTX 6000 Ada) |
| **GPU Memory (VRAM)**   | 12 GB or more ³                                               | 24 GB or more                                                          |
| **System Memory (RAM)** | 16GB                                                          | 32GB                                                                   |
| **Storage**             | 20GB free space                                               | 100GB free space                                                       |
| **Internet**            | Required for installation, authentication, and updates        | Required for installation, authentication, and updates                 |

<Info>
  **Important Notes:**

  ¹ Mac is not currently supported.

  ² Windows 10 is supported but may reach end-of-life before 2026 — Windows 11 is recommended.

  ³ GPUs with less than 12 GB VRAM may run but at unusable speeds.
</Info>

***

### Firewall & Network Configuration

If you're deploying Beeble Studio in a corporate or restricted network environment, ensure the following network requirements are met:

**Protocol & Port:**

* **HTTPS (TCP 443)** — All communication uses secure HTTPS over port 443

**Required Domains:**

| Domain                            | Purpose                                  |
| --------------------------------- | ---------------------------------------- |
| `app.beeble.ai`                   | Main application services                |
| `platform.modal.beeble.ai`        | Authentication and API services          |
| `release.beeble.ai`               | Software updates and AI model downloads  |
| `cdn.beeble.ai`                   | Public assets and static content         |
| `beeble-release.s3.amazonaws.com` | Software release storage                 |
| `clerk.beeble.ai`                 | User authentication                      |
| `accounts.beeble.ai`              | User authentication                      |
| `challenges.cloudflare.com`       | Bot protection challenges during sign-in |

<Tip>
  For enterprise deployments, add these domains to your firewall allowlist and
  ensure in/outbound HTTPS traffic on TCP port 443 is permitted.
</Tip>

***

### Using a Proxy Server

If your organization routes outbound traffic through an HTTP/HTTPS proxy — including TLS-inspecting corporate proxies — Beeble Studio requires the configuration below. Some components do not automatically follow the operating system's proxy settings, so each step is required for full functionality.

<Steps>
  <Step title="Launch the app with the proxy server flag">
    Start Beeble Studio with the `--proxy-server` flag pointing at your proxy. This routes the application's traffic and update checks through it.

    ```bash theme={null}
    beeblestudio --proxy-server="http://proxy.example.com:8080"
    ```

    <Note>
      On **Windows**, add the flag to the end of the **Target** field in the Beeble
      Studio shortcut's properties. On **Rocky Linux**, launch the app from a
      terminal with the flag.
    </Note>
  </Step>

  <Step title="Set the proxy environment variables">
    The local engine downloader and processing engine read the following environment variables. Set them in the environment before launching the app:

    ```bash theme={null}
    HTTPS_PROXY="http://proxy.example.com:8080"
    HTTP_PROXY="http://proxy.example.com:8080"
    REQUESTS_CA_BUNDLE="/path/to/your-proxy-certificate.pem"
    ```

    `REQUESTS_CA_BUNDLE` must point to your proxy's CA certificate. The certificate
    must be **PEM-encoded** — a text file that begins with
    `-----BEGIN CERTIFICATE-----`. If your certificate is in another format (for
    example, binary DER, `.p7b`, or `.pfx`), or you are unsure where to find it,
    ask your IT administrator for a PEM-formatted copy.
  </Step>

  <Step title="Trust the proxy certificate in Chrome">
    When you sign in, Beeble Studio opens the login page in your Chrome browser. If your proxy inspects HTTPS traffic, import your proxy's certificate so Chrome trusts it:

    **Chrome → Settings → Privacy and security → Security → Manage certificates →
    Custom → Installed by you → Trusted Certificates → Import**, then select your
    proxy certificate.
  </Step>
</Steps>

<Tip>
  Replace `http://proxy.example.com:8080` with your organization's proxy address,
  and set `REQUESTS_CA_BUNDLE` to the path of your PEM-encoded proxy certificate.
  Your IT or network administrator can provide both.
</Tip>

***

### Performance Benchmarks

Processing time per frame for VFX pass generation (PBR with alpha and depth) varies by GPU and resolution. See the benchmark below for reference.

| GPU Model   | 2K   | 4K   | 4K (4K Max Mode) |
| ----------- | ---- | ---- | ---------------- |
| L40S        | 1.9s | 2.9s | 6.6s             |
| A6000       | 3.0s | 4.0s | 12.2s            |
| RTX 5070 Ti | 2.0s | 3.0s | 8.0s             |
| RTX 4090    | 2.7s | 3.7s | 11.6s            |

<Info>
  **4K Max Mode** processes high-resolution footage at its native resolution (up
  to 4K). \ Requires at least 16 GB of GPU memory.
</Info>