Wan Pro Image-to-Video (fal)

Wan Pro is fal's premium Wan image-to-video endpoint - a higher-quality tier tuned for clean, sharp output - and this node exposes it with almost no dials to fuss over. You give it an image and a prompt, it animates the still. If you like Wan's motion quality (it's the community's reference for faces, bodies, and complex movement) but don't want to babysit a fifteen-setting local graph, this is the low-effort way to get it.

Cloud call, per-clip billing, no local GPU. Wan 2.2 is open, but Wan Pro is a hosted, enhanced endpoint you reach through fal.

How it works

Standard image-to-video: the still fixes the scene and subject, the prompt drives the motion, and Wan Pro generates a short clip. The "Pro" part is that fal handles the quality settings for you - resolution, steps, the lot - so the node stays minimal. That's the trade versus the Wan Animate/VACE nodes in this pack: less control, but a lot less to get wrong.

One habit worth carrying over from local Wan: for image-to-video, describe only the motion in your prompt, not the image. The picture is already set by your input frame. Restating "a woman in a red dress" when she's right there in the frame just muddies the motion instruction; "she turns and smiles as the camera pushes in" is what you want.

The inputs and outputs that matter

• prompt (required, multiline) - the motion and camera. Keep the scene description out of it.
• image (required, IMAGE) - the first frame. A crisp, high-quality still gives a much better clip.
• variations (optional, 1–10) - generate several takes at once and pick the best. Each one is billed.
• seed (optional, default 0) - set a fixed value to reproduce a result; the same seed with the same inputs re-rolls the same clip.
• enable_safety_checker (optional, default on).

Output is a STRING (a list - one URL per variation), each a link to a clip on fal. Download with the pack's Fal Save Media from URL node, or chain via *_direct_url. Save before the URLs expire.

How to install it

Ships in gokayfem's ComfyUI-fal-API pack. ComfyUI Manager: search ComfyUI-fal-API, install, restart. Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/gokayfem/ComfyUI-fal-API.git
pip install -r ComfyUI-fal-API/requirements.txt

Add your fal key (copy config.ini.example → config.ini, paste from fal.ai/dashboard/keys, or export FAL_KEY=your_key), restart.

Common issues & troubleshooting

Motion is weak or ignores the prompt. Describe the movement explicitly and leave the image description out. If it's still flat, a stronger action verb and a clearer input frame both help.

The clip wanders off your starting image. Image-to-video always trades some fidelity for motion. A sharper, less ambiguous still holds better; low-res or cluttered inputs drift.

I want more control over resolution/steps. Wan Pro deliberately hides those. If you need to tune them, the Wan Animate or VACE nodes in this pack (or a local Wan 2.2 setup) give you the knobs - at the cost of complexity.

Output won't preview. It's a URL string (or a list of them) - download via Fal Save Media from URL first.

Insufficient funds / credit lag. fal is pay-per-call and legit; top-ups have occasionally posted late. Wait for credits rather than repaying. fal errors print as text in ComfyUI.