LTX 2.3 Director Node Timeline Video Workflow

This ComfyUI workflow is designed for LTX 2.3 Director Node video generation, timeline-based prompt control, image-guided video creation, audio-video latent processing, and cinematic multi-segment video production. Instead of building a complex video workflow from many scattered prompt, image, and timing nodes, this workflow uses the Director Node logic to organize different visual instructions into a structured timeline.

The core idea is to treat video creation like directing a scene. Each segment can have its own start time, length, prompt, image reference, camera behavior, character focus, and dialogue description. This makes the workflow much easier to use for story-driven videos, character dialogue, cinematic short scenes, product demonstrations, music-video fragments, and multi-shot AI video generation.

The workflow is built around LTX 2.3, using ltx-2.3-22b-dev-dare-ties-distilled-1.1 as the main checkpoint, Gemma 3 12B with LTX text projection as the text encoder route, LTX23 audio VAE for audio latent processing, and the LTX 2.3 spatial upscaler v1.1 for latent video enhancement. It also includes LTXDirectorGuide, which is the key node for converting director-style timeline instructions into usable guide data for video generation.

A major advantage of this workflow is that it can combine image, text, and audio into one generation process. The Director Node can define multiple scene segments, while LTXVConditioning prepares the video conditioning at the selected frame rate. LTXVConcatAVLatent combines video latent and audio latent, and LTXVSeparateAVLatent separates them again after sampling. This makes the workflow suitable for audio-aware video generation and final video export with sound.

The workflow also includes a two-stage generation and refinement structure. The first pass creates the base video latent with director guidance. Then LTXVCropGuides and LTXDirectorGuide can continue refining the guided latent. LTXVLatentUpsampler is used with the spatial upscaler model to improve the final resolution and visual quality. This makes the workflow more suitable for polished outputs than a simple one-pass video render.

Sampling is handled through SamplerCustomAdvanced, BasicScheduler, CFGGuider, KSamplerSelect, and RandomNoise. This gives the workflow direct control over sampling steps, denoise strength, CFG behavior, random seed, and refinement stages. The included setup uses a compact step structure, making it practical for testing director-style segments without turning every render into a very heavy process.

Final output is handled by VAEDecode, LTXVAudioVAEDecode, CreateVideo, and VHS_VideoCombine. This means the workflow can decode generated video frames, recover the audio stream, combine image frames with audio, and export an MP4 file suitable for preview or publishing.

Main features:

- LTX 2.3 Director Node workflow

- Timeline-based video segment control

- Image + text + audio combined generation

- Segment-level prompt and camera control

- LTXDirectorGuide support

- LTXVCropGuides for guide refinement

- Audio-video latent concatenation and separation

- LTX 2.3 spatial upscaler v1.1 support

- Two-stage video generation and refinement

- CreateVideo and VHS_VideoCombine final export

- Suitable for cinematic scenes, dialogue clips, MV fragments, and AI story videos

Suggested workflow:

Prepare your source image, audio, and segment prompts first. Write each segment like a director instruction: who is speaking, what the character is doing, where the camera is pointing, what the lighting looks like, and how the shot should move. Keep mouth visibility clear if the scene includes dialogue. Use the Director Node timeline to control when each prompt or image reference becomes active. Start with short tests, confirm that the camera, character position, and timing are correct, then increase the segment length or refine the prompts.

🎥 YouTube Video Tutorial

Want to know what this workflow actually does and how to start fast?

This video explains what the tool is, how to launch the workflow instantly, and shares my core design logic — no local setup, no complicated environment.

Everything starts directly on RunningHub, so you can experience it in action first.

👉 YouTube Tutorial: https://youtu.be/wT9ob7rFONM

Before you begin, I recommend watching the video thoroughly — getting the full context helps you understand the tool faster and avoid common detours.

⚙️ RunningHub Workflow

Try the workflow online right now — no installation required.

👉 Workflow: https://www.runninghub.ai/post/2056305995550388226?inviteCode=rh-v1111

If the results meet your expectations, you can later deploy it locally for customization.

🎁 Fan Benefits: Register to get 1000 points + daily login 100 points — enjoy 4090 performance and 48 GB super power!

📺 Bilibili Updates (Mainland China & Asia-Pacific)

If you’re in the Asia-Pacific region, you can watch the video below to see the workflow demonstration and creative breakdown.

📺 Bilibili Video: https://www.bilibili.com/video/BV1FhLz6JEgJ/

☕ Support Me on Ko-fi

If you find my content helpful and want to support future creations, you can buy me a coffee ☕.

Every bit of support helps me keep creating — just like a spark that can ignite a blazing flame.

👉 Ko-fi: https://ko-fi.com/aiksk

💼 Business Contact

For collaboration or inquiries, please contact aiksk95 on WeChat.

🎥 YouTube 视频教程

想了解这个工作流到底是怎样的工具，以及如何快速启动？

视频主要介绍 工具定位、快速启动方法 和 我的构筑思路。

我们会直接在 RunningHub 上进行演示，让你第一时间看到实际效果。

👉 YouTube 教程： https://youtu.be/wT9ob7rFONM

开始前建议尽量完整地观看视频 —— 把握整体思路会更快上手，也能少走常见弯路。

⚙️ 在线体验工作流

现在就可以在线体验，无需安装。

👉 工作流： https://www.runninghub.ai/post/2056305995550388226?inviteCode=rh-v1111

打开上方链接即可直接运行该工作流，实时查看生成效果。

如果觉得效果理想，你也可以在本地进行自定义部署。

🎁 粉丝福利： 注册即送 1000 积分，每日登录 100 积分，畅玩 4090 体验 48 G 超级性能！

📺 Bilibili 更新（中国大陆及南亚太地区）

如果你在中国大陆或南亚太地区，可以通过下方视频查看该工作流的实测效果与构思讲解。

📺 B站视频： https://www.bilibili.com/video/BV1FhLz6JEgJ/

我会在 夸克网盘 持续更新模型资源：

👉 https://pan.quark.cn/s/20c6f6f8d87b

这些资源主要面向本地用户，方便进行创作与学习。