Kling Pro v1.6 Video Generation (fal)

KlingPro16 runs Kling's Pro v1.6 video endpoint inside ComfyUI, and its standout feature is the tail image - set a start frame and an end frame, and Kling generates the motion that connects them. That's the difference between hoping the model drifts somewhere good and telling it exactly where to land. It's the node for controlled transitions, deliberate camera moves, and clean loops.

Like everything in gokayfem's ComfyUI-fal-API pack, it runs on fal's servers: no local model, no VRAM, billed per clip. Kling itself remains one of the strongest closed video models around, so v1.6 Pro is a dependable workhorse even alongside the newer 2.6.

How it works

You give it a prompt and, optionally, a starting image and a tail image. With just a prompt it's text-to-video; add the start image and it's image-to-video; add the tail image too and Kling interpolates a path from the first frame to the last. fal runs it remotely and returns a string URL to the clip - save it before the link expires.

The inputs and outputs that matter

• prompt - describe the scene and how it should move.
• duration - 5 or 10 seconds. Ten costs more.
• aspect_ratio - 16:9, 9:16, or 1:1.
• image (optional) - the start frame; supplying it makes this image-to-video.
• tail_image (optional) - the end frame. This is the reason to pick this node over the plain Kling one: define where the clip finishes and let Kling fill the gap. Great for a controlled transition or, with the same image at both ends, a loop.

The single output is a STRING - the video URL.

How to install it

Via ComfyUI Manager: search ComfyUI-fal-API, install, restart. Manually:

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

No models download. Put a fal API key in config.ini under [API] (copy config.ini.example) or export FAL_KEY=your_key, restart, and find it under FAL/VideoGeneration.

Common issues & troubleshooting

• Auth error. FAL_KEY not set. Fix and restart.
• The transition looks abrupt. Start and end frames that are too different force the model to invent a lot of motion. Keep them related, or lengthen duration to give it room to move.
• The URL expired. fal result links don't last - route the STRING output into the pack's Save-Media-from-URL utility or download it right away.
• A run fails. The pack raises fal's real error message - content policy, quota, or a bad input. Read it.
• Cost. Each clip is a paid API call; 10 seconds costs more than 5. The result cache serves repeat identical calls free (force_rerun bypasses), and the spend guard caps a session.
• Windows Portable. Missing fal-client? python_embeded\python.exe -m pip install fal-client.

If you know where a shot should start and end and just need the in-between, this is the node. The tail image turns video generation from a gamble into something you can actually art-direct.