FAL Kling V3 Turbo Standard Image To Video (Soze)

Most of the FAL nodes in this pack are reference-to-video - a pile of images in, a prompt, a clip out. This one is deliberately simpler: exactly one image, exactly one prompt, one animated video. It's Kling V3 Turbo, the fast tier, wrapped for FAL. If you have a single great frame and you want it moving, this is the least-fiddly node you'll find.

The whole control surface

Three inputs, and only three required, which is the entire appeal:

• image - your first-frame reference. This is the frame the video starts from; Kling animates it forward. The node expects a single image, not a pile of references, so wire one IMAGE out of a Load Image node and be done.
• prompt - what happens next. "The camera slowly pushes in as the character turns to the window" - motion description beats static description here, because the image already carries the content.
• duration - a dropdown from 3 to 14 seconds. Longer clips cost more and risk more drift, so start at 5.

No aspect ratio, no resolution, no seed, no reference slots. That's the node. The "Turbo Standard" name means it's the standard-quality variant of the V3 Turbo line - fast for a video model, not free.

How it works

Same FAL machinery as the rest of the pack, just with a single upload: the image is converted and uploaded to FAL's CDN, the node POSTs to fal-ai/kling-video/v3/turbo/standard/image-to-video, waits synchronously, then downloads the mp4 into your output folder as a VIDEO object. Outputs are video (wire to a preview or saver), video_url (the CDN link), and status (a run summary).

Because it's single-image and single-shot, it slots into bigger workflows nicely: generate a still with your local model, feed it in here for the motion pass, then take the clip to your video editor. You don't even need to change how you build the rest of the graph.

Install and key

From 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 as an environment variable or in ComfyUI_Soze/config.ini:

[API]
FAL_KEY=your_key_here

Grab the key at fal.ai/dashboard/keys; the config file is gitignored, keep it that way.

Common issues

• Skips with no output. This one genuinely requires the image input - it's image-to-video, not text-to-video, and there's no reference fallback.
• The video ignores half your prompt. Keep motion in the prompt and leave visual description out; the image already fixes what things look like.
• Longer clips drift. Turbo is great for 5-second bursts; past 8 seconds expect the subject to wander and re-render instead of tweaking.