Creator builds narrative short film from LTX2.3 video and ACE-Step-1.5 music locally in ComfyUI

01Full local pipeline. Video synthesis, music generation, and editing all happened on consumer hardware. LTX2.3's open weights and ComfyUI's node-based interface made the entire process self-contained.

02Narrative structure. Most AI video showcases are single clips or tech demos. This piece has a premise, a character, and a three-act arc — the tunnel appears, the explorer descends, the underground world unfolds.

03Audio-visual sync. ACE-Step-1.5 generated the score in sync with the footage. Matching mood and pacing between two generative models in a local workflow is still rare in community projects.

04No cloud dependency. Running both models locally sidesteps API rate limits, usage caps, and per-generation costs. For creators building longer-form work, that's the difference between prototyping and finishing.

05ComfyUI as the hub. The workflow stitched together video generation, audio generation, and export without leaving the ComfyUI environment. Third-party nodes and custom pipelines are turning ComfyUI into a full production suite for local AI film work.

Creator builds narrative short film from LTX2.3 video and ACE-Step-1.5 music locally in ComfyUI

More in Community

What stands out