The video stage of our own pipeline, free: one ComfyUI node turns a picture into a 6-second clip with sound using LTX-2.5 (distilled, two stages: 8 steps at half resolution, x2 latent upscale, 3 steps at full resolution). About a minute per clip on a 24 GB card. 18+ only.

In the zip: Verboa Video (the whole image-to-video pipeline in one node), Verboa Act Prompt (writes the first-frame prompt and the motion prompt from a seed), Verboa Frame Sheet (a strip of frames to check before you post), and two workflows: verboa-video-i2v.json (animate any picture) and verboa-video-from-text.json (Act Prompt, then Verboa Image, then Verboa Video).

Set up: unzip into ComfyUI/custom_nodes/, restart ComfyUI, and put the five LTX-2.5 files from Lightricks/LTX-2.5 on Hugging Face in their models/ folders (the list is in the README inside the zip). For the first frame use Verboa Image 1.0 or any picture.

What we learned making clips: the first frame decides the clip (camera-facing, whole action in view, both faces visible); one woman and one man, one head each; keep "flushed" out of first-frame prompts; look at the frame strip before you post. LTX-2.5 has its own license: mark what you post as AI-generated.

Guide with examples: Verboa Video article. Prompt guide for the image model: Verboa Image prompt guide.