comfyui-wan21-i2v-cinematic-workflow

Wan 2.1 I2V — Cinematic Image-to-Video (ComfyUI Workflow)

A ComfyUI node graph (not model weights) for animating a still image into a cinematic clip with Wan 2.1 Image-to-Video (14B, GGUF Q3_K_S). Tested locally on an RTX 3090 — the demo clip linked in the full tutorial was generated with this exact workflow, using a FLUX.1 Dev keyframe as the starting image.

Recommended pipeline

1. Generate the keyframe with FLUX (or any image model); fix with inpainting if needed.
2. Load the image into this workflow and set motion_strength (start at 0.4-0.5).
3. Render first at 33 frames / 480p to validate motion before increasing resolution.

Requirements

• 12GB VRAM minimum.
• Nodes: WanVideoModelLoader, WanVideoTextEncode, WanVideoImageEncode, WanVideoSampler, WanVideoVAEDecode, VHS_VideoCombine (VideoHelperSuite custom node pack required).
• Time: 5-20 minutes per clip depending on GPU and resolution.

Full write-up

Full tutorial with troubleshooting and FAQ: https://comfylab.es/blog/workflows/wan-21-i2v-comfyui/ (Spanish) — English version at comfylab.dev if a translated version exists at upload time.

License

CC BY 4.0 — please attribute by linking back to the original article. This repo contains only the ComfyUI workflow JSON (node graph); it does not bundle or redistribute any model weights. Download Wan 2.1 separately per its own license.

Files

• wan-21-i2v-cinematic-comfylab.json — the workflow, load directly in ComfyUI.