one-take-vfx
Installation
SKILL.md
one-take-vfx
Run with python scripts/comfy_client.py --url $COMFY_URL run workflows/<file> --set NODE.INPUT=VALUE --out out/, and upload inputs first with comfy_client.py upload <file>. Before the first run, check the server with python scripts/preflight.py --url $COMFY_URL workflows/<file>. Node packs, model files, VRAM and weight licenses are listed in references/requirements.md. Use open-weights models only. LTX weights are under the Lightricks LTX license, which is not an OSI-approved open-source license.
Install (once)
Copy assets/onetake_vfx/ into ComfyUI/custom_nodes/onetake_vfx/. It is MIT, from Kainotomic/one-take-vfx, and provides OneTakeGuideMotion and OneTakeVFXPreset.
| Need | Workflow | Set |
|---|---|---|
| A preset effect | one_take_vfx_preset_node.api.json |
load_video.video, preset.preset (a name from references/presets/), noise.noise_seed |
| A custom effect prompt | one_take_vfx.api.json |
load_video.video, positive.text, empty_latent.length (8n+1), noise.noise_seed |
Rules
- Read the preset JSON before you change the prompt. It holds the tested prompt and strengths.
- Apply the effect to one hero shot, not every shot, and check continuity on the join into the next shot.
- The default working size is 768×448. Upscale afterwards with
video-upscale.