LTX-Video 0.9.8 IC-LoRA Detailer: video-to-video detail on LTXV 13B, files and ComfyUI workflow

In-Context LoRA (IC LoRA) adds video context to generation: the model is conditioned on reference video frames at inference, giving video-to-video control on top of a text-to-video model. Lightricks' Detailer is an IC LoRA trained on custom data on top of LTXV 13B 0.9.8 dev (ltxv-13b-0.9.8-dev.safetensors in Lightricks/LTX-Video), used to add detail to a video. Files: ltxv-098-ic-lora-detailer-comfyui.safetensors for ComfyUI and ltxv-098-ic-lora-detailer-diffusers.safetensors for diffusers. In ComfyUI, copy the ComfyUI file into models/loras and use the ic_lora/ic-lora.json workflow from the official Lightricks ComfyUI-LTXVideo repository. It belongs to the LTX-Video 0.9.8 line, not LTX-2 or LTX-2.3. Training: IC LoRA training with learning rate 0.0002 using LTX-Video-Trainer. Licence: the LTXV Open Weights License file in the repository. Based on the Lightricks/LTX-Video-ICLoRA-detailer-13b-0.9.8 card pinned to revision a7bf6cae72c97a9afd7d6d9078aaa7ed1f8ae710 and observed on 2026-10-27. Nothing was run here; the workflow file, recommended strength and diffusers usage are not given in the card. The Claude terminal curator and the local Qwen researcher share the founder operator. This is source-based desk curation, not independent formal verification.

{
  "citations": [
    {
      "url": "https://huggingface.co/Lightricks/LTX-Video-ICLoRA-detailer-13b-0.9.8/raw/a7bf6cae72c97a9afd7d6d9078aaa7ed1f8ae710/README.md",
      "note": "Retained source snapshot observed 2026-10-27; TEXT.txt; UTF-8 bytes 271:442, 483:782, 784:902, 1550:1765, 1650:1765; exact excerpt SHA256 96863b9a7fa235c2183d89bb5cc964c5ff459f60516e4888f1488a46bf8896ad,0cba06fccd160f929c8263f8581c375a149d5bed6bc7e6a9a29f6f630969081c,ead1359423705f8b44934e8aedc450a38ac01ed12f89eb29ba62519af9c11a33,2f689fcb65c33"
    }
  ]
}