FAL Kling O1 Standard Reference To Video (Soze)

If you want Kling O1 results in ComfyUI but the Pro tier's price per clip makes you wince, this is the node. It's the standard tier of the same Kuaishou model, wrapped for FAL, and for a lot of work - short clips, simple motion, iteration - you genuinely won't miss Pro. Same syntax, same controls, smaller bill.

The inputs

Identical shape to the Pro node, because it's the same class hitting fal-ai/kling-video/o1/standard/reference-to-video. Connect up to seven images to image_1–image_7 (or image_batch, capped at 7), cite them in the prompt as @Image1, @Image2, etc. duration is a 3–10 second dropdown, aspect_ratio is 16:9, 9:16, or 1:1. That's the whole control surface - there's no resolution picker, no seed, no audio option. If the clip is under 10 seconds and the motion is straightforward, this node covers it.

Where Standard genuinely wins

The obvious answer is cost, and it's the right one: Standard is the budget tier, so your FAL balance lasts longer when you're batching takes or storyboarding. But there's a second, less obvious argument - for simple shots, Standard's faster turnaround means more iterations per session, and iterating beats polishing the wrong idea with the premium model. The Pro tier earns its money on complex motion, fast camera moves, and multi-subject scenes. A character standing, walking, or turning? Standard will likely deliver that at a fraction of the price.

So the workflow pattern that works: rough everything out on Standard, then spend the Pro credits only on the final take that has to be clean. The two nodes share the same citation format, so swapping them in is a matter of deleting one node and dropping in the other - the prompts and image wiring transfer as-is.

How it works

Same FAL machinery as the rest of the Soze pack: reference images upload to FAL's CDN, the node POSTs to the O1 Standard endpoint, waits synchronously, then downloads the mp4 into your output folder as a VIDEO object. Outputs are video, video_url, and status. Nothing runs on your GPU beyond file handling.

Install and key

Part of the Soze pack ("Quality of Life Nodes for ComfyUI"). ComfyUI Manager → search "Soze", or:

cd ComfyUI/custom_nodes
git clone https://github.com/SozeInc/ComfyUI_Soze.git
pip install -r ComfyUI_Soze/requirements.txt

Restart, then set FAL_KEY (environment variable or ComfyUI_Soze/config.ini):

[API]
FAL_KEY=your_key_here

Key at fal.ai/dashboard/keys; the config file is gitignored, leave it that way.

Common issues

• Skips with no output. Standard is still reference-to-video; connect at least one image.
• Motion quality disappoints on complex shots. That's the tier boundary, not a bug - this is where you step up to Pro for the final render.
• Long waits. Synchronous call, and video generation is minutes, not seconds. Queue it and go do something else.