KlingAI 2.1 Standard

Kling 2.1 Standard is the value tier of the 2.1 generation, and like its Pro sibling it's image-to-video-first: frameImages and positivePrompt are both required. Same model family, same required-input shape, lower resolution ceiling, cheaper per clip. If you've got a frame and you want motion without paying flagship prices for every draft, this is the node that makes iteration affordable.

Why you'd reach for it

Volume. The 2.1 line brought real improvements in motion and subject consistency, and Standard gives you that improvement at a price you can afford to spend in a loop. Draft, render, look, redraft - with a value tier, that loop doesn't make your wallet flinch. When one of those drafts finally hits, that's when you swap the Standard node for the Pro version and re-render the keeper. It's a two-node workflow that ends up cheaper than gambling on Pro from the start.

How it works

The pack's standard generated node: your frameImages tensor is base64'd to PNG data URIs, bundled with the prompt, and POSTed to the Runware REST API via the runware SDK. The clip comes back as a native VIDEO object and the cost prints on the title bar. Nothing renders on your machine.

Inputs that matter

• frameImages (IMAGE) - required. First frame from Load Image or any IMAGE output.
• positivePrompt - required. Motion and scene changes, not a re-description of the still.
• duration - 5 or 10 seconds.
• width / height - 1024×1024 default. Don't overshoot: on the Standard tier, extra pixels buy little and cost real credits.
• CFGScale - 0.5 default; the model's default guidance behavior is what you'll mostly use here.
• numberResults (1–4), outputFormat (MP4/WEBM/MOV), outputQuality (95).

Output: one video socket (VIDEO type).

Install and the API key

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

Restart ComfyUI (or Runware via ComfyUI Manager). Key from the Runware dashboard → Settings → Runware API key, RUNWARE_API_KEY, or runware auth login.

Gotchas

• frameImages wants a tensor - a filename or URL will fail with a confusing error. Load Image first.
• No mask input: whatever's in the frame gets animated, so crop or composite before you send it if you want control.
• Closed model, per-request pricing: Kuaishou's weights, rented by the clip, cost on the node.
• The VIDEO output needs current ComfyUI; older builds fall back to a saved file path.

Standard isn't the bragging-rights tier - it's the "do the work" tier. Pair it with Pro as a finish pass and you've got a genuinely smart pipeline.