Descriptions:
Can AI video move beyond simple “Image-to-Video” and start telling logical stories? 🎬 Welcome to LTX 2.3 MSR V2 (Multiple Subject Reference)! This isn’t just an incremental update; it’s a revolutionary way to organize your AI assets into a coherent event.
In this tutorial, I’ll show you how to use the LiconMSR node to package characters, props, and backgrounds into a “Visual Memory” sequence. We’ll test the limits of its “Event Logic” (from seed to bloom) and see how it performs in complex Multi-Shot scenarios where character identity usually fails.
🔥 What you’ll master in this tutorial:
• 🧠 The MSR V2 Logic: How to turn multiple reference images into a chronological event.
• ⚙️ ComfyUI Workflow Breakdown: Mastering the LiconMSR node and the new PromptRelayEncode.
• 🎨 Structured Prompting: Why you need to describe your “Reference Sheet” and “Target Video” separately for best results.
• 🧪 Stress Tests: From 2-character interactions to the extreme 5-reference image challenge.
• 🎞️ Cinematic Consistency: How to recover character identity (hair, clothes, face) even after a shot cut.
#runninghub #LTX23 #MSRV2 #ComfyUI #AIVideo #CharacterConsistency #MultipleSubjectReference #LiconMSR #PromptRelay #AIAnimation #AIFilmmaking #GenerativeAI #VideoLogic #RunningHub #AITutorial #VideoGeneration
————————————————————————————————————————
resource
https://github.com/liconstudio/ComfyUI-Licon-MSR
https://huggingface.co/LiconStudio/LTX-2.3-Multiple-Subject-Reference
online workflow
1：LTX-2.3-MSR_V2_Sample_workflow
https://www.runninghub.ai/post/2077277675710054401/?inviteCode=kol02-rh021
Runninghub Fan Benefit: Register to get 1,000 RHB
https://www.runninghub.ai/?utm_source=kol02-RH021
workflow and prompt engineering download：
https://github.com/amao2001/ganloss-latent-space/blob/main/workflow/2026-07-17%20LTX-2.3-MSR_V2_Sample_workflow%20.json
————————————————————————————————————————
