ID-LoRA

ID-LoRA / ID-LoRA-LTX2.3-ComfyUI

Public

Custom ComfyUI node for generating videos with audio-visual identity based on a reference voice and image

id-lora.github.io audio-visual-generation comfyui-custom-node video-generation
45
7
100% credibility
Found Mar 23, 2026 at 45 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

ComfyUI custom nodes for generating talking-head videos that transfer speaker identity from a reference audio using ID-LoRA-2.3 one-stage or two-stage pipelines.

How It Works

1
👀 Discover the fun tool

You find this exciting ComfyUI add-on that lets you create custom videos of people talking with any voice you choose.

2
💻 Add it to your playground

Follow the simple guide to bring this tool into your ComfyUI workspace so everything is ready to play.

3
📥 Gather creative pieces

Download the special artwork and sound kits – they take up space but make the magic happen on your powerful computer.

4
Pick your video style
⚡
Quick mode

Go for standard quality to get results fast.

⭐
Fancy mode

Choose high-quality for sharper, more detailed videos.

5
🖼️ Choose face and voice

Select a photo for the person's face and a short voice clip to copy the speaking style.

6
✍️ Describe the scene

Write what the person looks like, what they say, and any extra sounds you want.

7
▶️ Hit create

Click the button and watch as your video comes to life step by step.

🎉 Enjoy your video

Save your personalized talking video with the perfect voice and movements, ready to share.

Sign up to see the full architecture

6 more

Sign Up Free

Star Growth

See how this repo grew from 45 to 45 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is ID-LoRA-LTX2.3-ComfyUI?

This Python-based ComfyUI extension adds custom nodes to generate talking-head videos that transfer a speaker's voice identity from a reference audio clip onto a new visual scene, powered by LTX-2.3 and ID-LoRA-2.3 models. Users supply a first-frame image, reference voice sample, and structured prompts—like [VISUAL] for scenes, [SPEECH] for dialogue, [SOUNDS] for audio style—to output synced video+audio clips up to 1024x1024 at 25fps. It solves the hassle of scripting audio-visual personalization by plugging into ComfyUI's comfyui custom nodes folder via standard comfyui custom model path symlinks.

Why is it gaining traction?

Unlike standalone ID-LoRA scripts, it offers drag-and-drop ComfyUI nodes for one-stage (fast, 512px) or two-stage (HQ 2x upsampling) pipelines, with built-in quantization (int8/fp8) to fit 24GB GPUs and example workflows for instant comfyui custom nodes github testing. Developers love the auto-resolving comfyui custom nodes list, prompt encoding that frees VRAM early, and seamless SaveVideo output—no more comfyui custom nodes import failed errors after setup. The custom comfyui manager-like dropdowns for checkpoints, LoRAs, and text encoders make iteration feel native.

Who should use this?

ComfyUI workflow builders crafting personalized avatars for apps, AI video editors syncing custom voices to faces, or researchers prototyping identity-driven talking heads. Ideal for those with NVIDIA GPUs (24GB+ VRAM) handling 67GB+ models, following a comfyui custom node tutorial for symlinks and pip installs.

Verdict

Grab it if you're deep in ComfyUI custom workflows and have the hardware—docs and templates are solid for quick starts, despite 45 stars and 1.0% credibility signaling early maturity. Skip for lighter setups; model downloads and VRAM demands aren't beginner-friendly yet.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.