Loop ForgeCharacter, prop and location
Works, from LTX 2.5 Creative Lab LoRAs, tested locally
- LoRA
- Ingredients at 1.4
- Output
- 960×544, 121 frames at 24fps
- Sampling
- 8 steps, euler_ancestral_cfg_pp
- Seed
- 42
- Generated in
- 12 min
- Peak VRAM
- 11.4 GiB
- Ran on
- RTX 3060, 12GB (local)
What was wired in
Reference sheet
Rooftop sheetIngredients sheet: character, camera and rooftop panels on blackDownload PNG0.7 MB
The woman, her yellow camera and the rooftop each come from a separate panel of the sheet, and they end up together in a new moving shot.
Reference sheet: a portrait photo of a young woman with shoulder-length dark hair streaked pink and blue, large gold hoop earrings, a small tattoo on her neck and a beige knitted sweater; a product photo of a vintage yellow 35mm film camera with a black strap; a plain photo of an empty rooftop terrace with a brick parapet, potted olive trees and a city skyline. Generated video: the same woman is standing on the rooftop terrace at golden hour, lifting the yellow film camera to her eye, taking a photo of the skyline and then lowering it with a small smile toward the camera, in a static medium shot with a light breeze and distant traffic below.
worst quality, inconsistent motion, blurry, jittery, distorted
Notes
The sheet follows the Lightricks demo layout: a 1536×896 black canvas, 16 px gutters and plain photo panels. Keeping to that sizing and those margins gave me better results.
Run it yourself
Download everything for this test (.zip, 0.7 MB). It holds the workflow, every input it loads and a README. Copy the inputs into ComfyUI/input, put the models below in place and drag the workflow onto ComfyUI.
| File | Put it in | Size |
|---|---|---|
| ltx-2.5-22b-distilled-transformer-comfy-int8-convrot.safetensors gated | models/diffusion_models | 21.5 GB |
| gemma4-12b-with-proj-ltx-2.5-comfy-int8-convrot.safetensors gated | models/text_encoders | 15.4 GB |
| ltx-2.5-video-vae-bf16.safetensors gated | models/vae | 1.5 GB |
| ltx-2.5-audio-vae-bf16.safetensors gated | models/vae | 364.9 MB |
| ltx-2.5-22b-ic-lora-ingredients-0.9.safetensors gated | models/loras | 1.3 GB |
The Lightricks repositories (the LTX 2.5 base model and the LoRAs) are gated. Log in to Hugging Face and accept the terms on each repository page before downloading.
Custom nodes: ComfyUI-LTXVideo at commit 3bf3ca6. Everything else is part of ComfyUI 0.37.0 (commit 0f74f7fb, frontend 1.53.6).