<div data-youtube-video="https://youtu.be/gWAvmRaJ9gY"><iframe src="https://www.youtube.com/embed/gWAvmRaJ9gY" width="853" height="480" allowfullscreen></iframe></div> 
 
This LTX 2.5 workflow turns an existing video into a higher-resolution generative redraw. It is designed for creators who need more than conventional interpolation or sharpening: the input is enlarged, encoded, refined through an LTX video generation pass, decoded in tiles, and exported while retaining the source timing. 
 
The active graph loads the LTX 2.5 22B distilled BF16 transformer, the Gemma 4 12B LTX text encoder in INT8 ConvRot format, the LTX 2.5 video VAE, and the LTX IC detailer LoRA at strength 1.0. A configurable RTX video-upscale stage prepares the larger frame sequence before tiled VAE encoding. The LTX looping sampler and Euler sampler then perform the generative detail pass, followed by spatiotemporal tiled VAE decoding and a final video export route. 
 
The workflow is packaged as a ready-to-run ComfyUI graph for users who want to compare generative enlargement against conventional video scaling. Keep the original source, prompt, and duration unchanged when comparing versions so that visible differences come from the workflow structure rather than from changed creative inputs. 
 
Main features: 
 
- Video-to-video generative high-resolution redraw 
- LTX 2.5 22B distilled transformer route 
- Gemma 4 12B LTX text encoding in INT8 ConvRot format 
- LTX IC detailer LoRA enabled at strength 1.0 
- Configurable RTX spatial enlargement before regeneration 
- Tiled VAE encode and spatiotemporal tiled decode 
- Euler sampling with the LTX looping sampler 
- Final 30 fps video creation and export 
 
Suggested workflow: 
 
Load a short source video with stable motion and a readable main subject. Choose the enlargement target first, then describe only the details that should be regenerated. Keep identity, clothing, lighting, and camera instructions explicit. Test a short section before processing a long clip, because stronger detail prompts can also reinterpret small textures and background elements. 
 
RunningHub Workflow 
 
Try the workflow online right now - no installation required. 
Workflow: https://www.runninghub.ai/post/2088895590487244801?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 &amp; Asia-Pacific) 
 
Watch the workflow demonstration and creative breakdown here: 
Bilibili Video: https://www.bilibili.com/video/BV1kZbQ6xEMg/ 
 
Support Me on Ko-fi 
 
Ko-fi: https://ko-fi.com/aiksk 
 
Business Contact 
 
For collaboration or inquiries, please contact aiksk95 on WeChat. 
 
打开下方链接即可在线体验，无需安装。 
工作流：https://www.runninghub.ai/post/2088895590487244801?inviteCode=rh-v1111 
 
如果觉得效果理想，也可以在本地进行自定义部署。 
 
粉丝福利：注册即可领取 1000 积分，每日登录再领 100 积分，体验 4090 和 48 GB 大显存性能。 
 
B站视频：https://www.bilibili.com/video/BV1kZbQ6xEMg/ 
 
我会在夸克网盘持续更新模型资源： 
https://pan.quark.cn/s/07bdc81784ce

LTX 2.5 Generative Video Upscale

<div data-youtube-video="https://youtu.be/gWAvmRaJ9gY"></div> 
 
This LTX 2.5 workflow turns an existing video into a higher-resolution generative redraw. It is designed for creators who need more than conventional interpolation or sharpening: the input is enlarged, encoded, refined through an LTX video generation pass, decoded in tiles, and exported while retaining the source timing. 
 
The active graph loads the LTX 2.5 22B distilled BF16 transformer, the Gemma 4 12B LTX text encoder in INT8 ConvRot format, the LTX 2.5 video VAE, and the LTX IC detailer LoRA at strength 1.0. A configurable RTX video-upscale stage prepares the larger frame sequence before tiled VAE encoding. The LTX looping sampler and Euler sampler then perform the generative detail pass, followed by spatiotemporal tiled VAE decoding and a final video export route. 
 
The workflow is packaged as a ready-to-run ComfyUI graph for users who want to compare generative enlargement against conventional video scaling. Keep the original source, prompt, and duration unchanged when comparing versions so that visible differences come from the workflow structure rather than from changed creative inputs. 
 
Main features: 
 
- Video-to-video generative high-resolution redraw 
- LTX 2.5 22B distilled transformer route 
- Gemma 4 12B LTX text encoding in INT8 ConvRot format 
- LTX IC detailer LoRA enabled at strength 1.0 
- Configurable RTX spatial enlargement before regeneration 
- Tiled VAE encode and spatiotemporal tiled decode 
- Euler sampling with the LTX looping sampler 
- Final 30 fps video creation and export 
 
Suggested workflow: 
 
Load a short source video with stable motion and a readable main subject. Choose the enlargement target first, then describe only the details that should be regenerated. Keep identity, clothing, lighting, and camera instructions explicit. Test a short section before processing a long clip, because stronger detail prompts can also reinterpret small textures and background elements. 
 
RunningHub Workflow 
 
Try the workflow online right now - no installation required. 
Workflow: https://www.runninghub.ai/post/2088895590487244801?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 &amp; Asia-Pacific) 
 
Watch the workflow demonstration and creative breakdown here: 
Bilibili Video: https://www.bilibili.com/video/BV1kZbQ6xEMg/ 
 
Support Me on Ko-fi 
 
Ko-fi: https://ko-fi.com/aiksk 
 
Business Contact 
 
For collaboration or inquiries, please contact aiksk95 on WeChat. 
 
打开下方链接即可在线体验，无需安装。 
工作流：https://www.runninghub.ai/post/2088895590487244801?inviteCode=rh-v1111 
 
如果觉得效果理想，也可以在本地进行自定义部署。 
 
粉丝福利：注册即可领取 1000 积分，每日登录再领 100 积分，体验 4090 和 48 GB 大显存性能。 
 
B站视频：https://www.bilibili.com/video/BV1kZbQ6xEMg/ 
 
我会在夸克网盘持续更新模型资源： 
https://pan.quark.cn/s/07bdc81784ce

Files with SHA256: 767443b2a9914ccde4a8f2ba2e0dde71164dc6f0238a01b48271dc087aadae31