/quickdesign

When to use

• Budget-tight job, simple animation, no spoken voiceover
• Non-UGC b-roll: product rotation, atmospheric loop, abstract motion
• User explicitly named Kling
• Voice continuity / multi-product reference fidelity is NOT required

Don't use for spoken-script UGC — Kling's TTS is weaker than Seedance's native voice and there's no --reference-audio continuity. For talking-avatar work always start at Seedance R2V.

Hard facts (live)

quickdesign cost kling-3-pro -d 5
quickdesign video models | jq '.data[] | select(.slug=="kling-3-pro")'

• Duration grid: 3..15s (per-second lookup, see quickdesign cost kling-3-pro -d <n>)
• Cost: duration_lookup — discrete cost per duration; check before submitting.
• Reference grammar: NONE. Single --image input only.
• Audio: yes via generate_audio, but voice character generally weaker than Seedance for speech.

Compact prompt skeleton

<scene description>. <action / motion verbs OK here, unlike Seedance>.
<aspect ratio> format.

Camera-motion verbs ("slowly zooms in", "pans left", "tracking shot") work better here than in Seedance — Kling produces cleaner cinematic motion when explicitly directed.

Sibling models in the Kling family

• kling-3-standard — cheaper, simpler animation. Use for loops where Pro's quality bump isn't worth it.
• kling-2.6-pro, kling-2.1-pro, kling-2.1-standard — older versions, retained for users with prior workflows. Default to kling-3-* for new work.
• kling-o1-edit — fixed-cost edit operation (64cr), niche use.

Compare via quickdesign cost --category video | grep kling.

Gotchas

1. No audio_urls continuity — every segment generates a fresh voice. Multi-segment Kling is voice-incoherent unless you use TTS + lipsync external pipeline.
2. No @Image1 reference grammar — descriptive prompts carry more weight than with Seedance R2V. Be more verbal.
3. Subtitle / music defaults are less aggressive than Seedance, but still add No music score. + No subtitles. to be safe.

Cross-references

• For talking-script UGC use Seedance R2V instead → ./seedance-2.0-r2v.md
• Voice continuity strategies → ../references/voice-continuity.md