LatentSync Lip Sync Guide for Low VRAM ComfyUI Users

‍LatentSync is one of the best lip sync tools available today — and now, you can run it smoothly on low VRAM systems using ComfyUI. This powerful tool generates realistic lip movements without needing masks, making it perfect for creators seeking professional-quality sync without requiring high-end GPUs.

💻 What Is Promptus Studio Comfy?

While ComfyUI is a robust open-source framework, many users prefer to use it through Promptus Studio Comfy (PSC) — a browser-based, cloud-powered visual AI platform built on ComfyUI.

Key Features of Promptus Studio Comfy:

• 🔧 No-code workflow builder via CosyFlows
• 🔗 Real-time collaboration
• 🧠 Built-in access to Gemini Flash, HiDream, Hunyuan3D
• 💬 Discord integration
• 📦 Workflow publishing
• 🖼️ Multi-modal generation (text, image, video)
• ⚡ Distributed GPU compute for fast, high-res outputs

🔧 Installing LatentSync on ComfyUI

Step-by-Step Setup:

1. Open your ComfyUI folder → go to the custom nodes subfolder.
2. Type cmd in the path bar to open a command window.
3. Go to the LatentSync GitHub repo and copy the code URL.
4. Run the following command in CMD:

git clone <LatentSync-URL>
cd LatentSync


5. Open the requirements.txt file and remove the torch / pytorch lines (ComfyUI already includes them).
6. Save the file and run:

pip install -r requirements.txt


📦 Installing Additional Dependencies

1. Still in the custom nodes folder, open a new command window.
2. Go to the aceNodes GitHub repo, copy the code URL.
3. In CMD, run:

git clone <aceNodes-URL>
cd aceNodes


4. Again, open requirements.txt, remove torch, save, and run:

pip install -r requirements.txt


⚡ Using LatentSync with ComfyUI

Once installed:

1. Run ComfyUI from your main folder.
2. Download the LatentSync workflow and load it into your canvas.
3. Upload:
   • 🗣️ An audio file (.wav)
   • 🎥 A video file (.mp4)
4. Make sure both files are the same length.
5. Hit queue to begin the sync process.

⚠️ The first run may auto-download required models.

☁️ Running LatentSync on Kaggle (Cloud GPU)

Cloud Setup Instructions:

1. Open the official Kaggle notebook and click "Edit My Copy".
2. In the Kaggle runtime settings:
   • Enable GPU
   • Turn on Internet
   • Set Persistence to: Files and Variables
3. Start the session, then:
   • ▶️ Run the first cell (for installation)
   • ▶️ Then run the Pingy cell
4. Click on the Pingy local URL to launch ComfyUI in your browser
5. Load your workflow, upload audio/video files, and hit generate!

🌟 Conclusion

LatentSync is a game-changer for lip sync AI — especially for creators with low VRAM systems. Whether you're running it locally or in the cloud (via Kaggle), it delivers smooth, natural, high-quality lip sync video generation.

🚀 For a Streamlined Experience: Try Promptus

Promptus Studio Comfy combines:

• 🧩 ComfyUI’s power
• 🖱️ Drag-and-drop usability
• ⚡ High-speed GPU rendering
• 🌐 Web + Desktop support

➤ Sign up today at: https://www.promptus.ai

Start creating expressive, synced videos without technical hassle — ideal for:

• 🎬 Branded content
• 🎨 Animated stories
• 🧙 Concept art pipelines

Let creativity lead, and let the AI do the heavy lifting. 💡