KlingAI 1.0 Standard

Kling 1.0 Standard is the cheap seat at the same table. It's the value tier of Kuaishou's first big video model, and in the Runware pack that means a real Kling render with a smaller credit bill than the Pro variant. If you've never run a cloud video node and want to see what the fuss is about before committing, this is the node to burn your first few cents on.

What's actually different

Pro and Standard share the same 1.0 architecture, so your prompt, your workflow, and your wiring are identical - you're just choosing a tier. The Standard model resolves at a lower ceiling (think 720p-class output vs the Pro tier's 1080p) and prices out cheaper per clip. For quick motion tests, storyboards, or anything that's going to be a thumbnail anyway, Standard is the sensible default. For hero shots you'd rather have, the Pro node is two search-box entries away.

How it works

Same pack machinery as every Kling node: the class is generated from Runware's model catalog, and on queue your positivePrompt (plus any frameImages tensor, which gets base64'd to PNG data URIs) is sent to the Runware REST API. The rendered clip is downloaded and returned as a native ComfyUI VIDEO object. Nothing runs on your machine - that's the whole point of the pack.

Inputs that matter

Only positivePrompt is required. From there:

• frameImages (IMAGE) - optional. Feed an image to turn this into image-to-video; skip it for pure text-to-video.
• duration - 5 or 10 seconds. The longer clip costs more per run, and at Standard resolution you'll feel the value trade-off.
• width / height - 1024×1024 by default.
• CFGScale - 0.5 default; negativePrompt is ignored below ~1, so only bother writing one if you raise guidance.
• numberResults - 1 to 4 variations per run.
• outputFormat (MP4/WEBM/MOV) and outputQuality (95 default) control the container and encoder.

Output is one video socket (VIDEO type) into a preview/save node.

Install and the API key

If you already have the Runware pack, this node is just sitting in your Runware/Video/klingai menu. Otherwise:

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

Restart ComfyUI (or search Runware in ComfyUI Manager). Get a key from the Runware dashboard, then set it in Settings → Runware API key, via RUNWARE_API_KEY, or with runware auth login.

Gotchas

• No key, no run - you'll get the pack's one-line "No Runware API key" error.
• This is paid inference; the node's title bar shows the cost of each run so the pricing isn't a mystery.
• frameImages expects a real IMAGE tensor from a Load Image node, not a file path.
• Remember the model itself is closed - Kling is Kuaishou's, API-only. You're renting the inference, not owning the weights.

Standard-tier Kling 1.0 is the budget option of an old model. It won't win any beauty contests against the 1.6 or 2.x nodes, but as a cheap, low-risk introduction to driving cloud video from ComfyUI, it does its job.