# Transform into anyone using this crazy AI tool

Welcome to the easiest guide you'll find for installing **Deep Live Cam**, a fun and impressive live deepfake tool. Even if you’re not technical, don’t worry, we’ve got your back.

## What is Deep Live Cam?

It's a tool that lets you transform your face in real time using AI magic. Think FaceSwap, but live. You can prank your friends or just explore the tech, it’s all local (on your computer), so no data is sent anywhere. Your secret's safe.

---

## What You'll Need First

Before installing, let's get the essentials out of the way. It’s like setting the table before a feast.

### 1\. Install Python (3.10 is required)

Head over to [python.org](https://www.python.org/downloads/release/python-3100/) and download Python 3.10.

> Make sure you check the box that says **"Add Python to PATH"** during installation.

### 2\. Install Git

Grab it from [git-scm.com](https://git-scm.com/) and install like you would any other app.

### 3\. Install `ffmpeg`

This tool handles video magic.

* Press `Windows + R`, type `powershell`, then hit enter.
    
* Paste this in and press Enter:
    

```bash
iex (irm ffmpeg.tc.ht)
```

This command:

* `iex` = run the downloaded command as a script
    
* `irm` = download a script from the web (shorthand for `Invoke-RestMethod`)
    
* The rest points to a safe script to auto-install ffmpeg on Windows.
    
* after the script runs you'll get three options and you need to choose number 3
    
* still stuck? refer this video: [https://www.youtube.com/watch?v=OlNWCpFdVMA](https://www.youtube.com/watch?v=OlNWCpFdVMA)
    

### 4\. Install Visual Studio 2022 Runtimes (Windows only)

Download from: [Microsoft Visual C++ Redistributable](https://learn.microsoft.com/en-us/cpp/windows/latest-supported-vc-redist?view=msvc-170)

> Choose the version matching your system (x64 for most users).

---

## Clone & Set It Up

### 1\. Open Terminal / Command Prompt and paste the commands:

```bash
git clone https://github.com/hacksider/Deep-Live-Cam.git
```

```bash
cd Deep-Live-Cam
```

### 2\. Download the Models

Download these two files:

* [`GFPGANv1.4`](https://huggingface.co/hacksider/deep-live-cam/resolve/main/GFPGANv1.4.pth)
    
* [`inswapper_128_fp16.onnx`](https://huggingface.co/hacksider/deep-live-cam/resolve/main/inswapper_128_fp16.onnx)
    

Put them into the **models** folder inside the cloned Deep-Live-Cam directory. No need to unzip or rename.

### 3\. Set Up Virtual Environment (Optional, but Smart)

Helps prevent your computer from becoming a dependency mess.

#### For Windows:

```bash
python -m venv venv
```

```bash
venv\Scripts\activate
```

```bash
pip install -r requirements.txt
```

#### For macOS (Apple Silicon):

```bash
brew install python@3.10
```

```bash
brew install python-tk@3.10
```

```bash
python3.10 -m venv venv
```

```bash
source venv/bin/activate
```

```bash
pip install -r requirements.txt
```

> If you get `_tkinter` errors, reinstall tkinter using `brew reinstall python-tk@3.10`

#### For Linux:

```bash
python3.10 -m venv venv
```

```bash
source venv/bin/activate
```

```bash
pip install -r requirements.txt
```

### 4\. Fix Missing Module Error (pygrabber)

You might see an error like `ModuleNotFoundError: No module named 'pygrabber'` Just run:

```bash
pip install pygrabber
```

---

## Time to Run It!

You can run the tool with this command:

```bash
python run.py
```

This downloads additional files (~300MB), so have your Wi-Fi ready.

## GPU Acceleration (If You Want Speed)

### CUDA (Nvidia Only)

**Only try this if you're feeling adventurous and have an Nvidia GPU.**

**Steps:**

1. Install CUDA Toolkit 11.8 from [Nvidia](https://developer.nvidia.com/cuda-11-8-0-download-archive)
    
2. You also need a full Visual Studio 2022 installation with C++ workload enabled.
    

> It didn't work well on a GTX 3050 for me. YMMV.

Then run:

```bash
pip uninstall onnxruntime onnxruntime-gpu
```

```bash
pip install onnxruntime-gpu==1.16.3
```

```bash
python run.py --execution-provider cuda
```

### Apple Silicon (M1/M2/M3)

```bash
pip uninstall onnxruntime onnxruntime-silicon
```

```bash
pip install onnxruntime-silicon==1.13.1
```

```bash
python3.10 run.py --execution-provider coreml
```

### Windows (DirectML)

```bash
pip uninstall onnxruntime onnxruntime-directml
```

```bash
pip install onnxruntime-directml==1.15.1
```

```bash
python run.py --execution-provider directml
```

### Intel (OpenVINO)

```bash
pip uninstall onnxruntime onnxruntime-openvino
```

```bash
pip install onnxruntime-openvino==1.15.0
```

```bash
python run.py --execution-provider openvino
```

---

## Troubleshooting Tips

* If your models don’t load: double-check they’re in the right **models/** folder.
    
* If Python 3.10 isn’t being used:
    
    * On macOS, always use `python3.10`, not `python`
        
* Conflicting Python versions?
    

```bash
brew list | grep python
brew uninstall --ignore-dependencies python@3.11 python@3.13
brew cleanup
```

---

## Final Notes

* The tool works fine with just CPU, though it’ll be slower.
    
* GPU acceleration is cool, but not mandatory.
    
* If you hit a wall, don’t panic—90% of the time it’s a missing install.
    

> Now go deepfake responsibly.

## Credits

Deep Live Cam is created and maintained by [hacksider on GitHub](https://github.com/hacksider/Deep-Live-Cam?tab=readme-ov-file).