All skills
anthropics avatar

/quickdesign

@710a472

Use the `quickdesign` CLI to generate AI media — UGC promo videos, image edits, product creatives, video upscales — through Seedance, Kling, Sora2, Nano Banana, and GPT Image. Invoke this skill whenever the user asks for a talking-avatar video, multi-segment ad / promo / explainer, image edit (object swap, angle change, state change), product photoshoot, or video upscale via QuickDesign.

Use this Skill: https://skilld.dev/gh/anthropics/claude-plugins-community/quickdesign

This session only. Nothing lands on disk.

modelstopaz-video-upscale.md

≈646 tokens on demand. Your agent reads this file only when SKILL.md points to it.

When to use

  • Final concatenated video is too low-res for the delivery channel (1080p reel posted to a platform that prefers 4K masters)
  • Source footage is older / lower res and you want to clean it up before publish
  • Text-heavy / detail-heavy frames need extra crispness

Don't use as a substitute for generating at 1080p in the first place. Seedance R2V at 1080p produces a cleaner master than 720p + Topaz upscale, in most cases.

Hard facts (live)

quickdesign cost topaz-video-upscale -d 24 -r 4k
quickdesign cost topaz-video-upscale -d 24 -r 4k --fps60
quickdesign video models | jq '.data[] | select(.slug=="topaz-video-upscale")'
  • Cost: per_second × resolution. 720p=1cr/s, 1080p=2cr/s, 4k=8cr/s.
  • 60fps multiplier: ×2 if --fps60. Use ONLY if downstream playback is genuinely 60fps; most social platforms are 30fps and the bump is wasted.
  • Sibling: bytedance-video-upscale is similar with a different cost ladder + a pro_multiplier. Compare via quickdesign cost --category video_upscale.

Pipeline placement

Always run after the final concat, not per-segment:

# generate segments at native resolution
quickdesign video generate ... -o seg1.mp4
quickdesign video generate ... -o seg2.mp4

# concat (mux-only, no quality loss)
ffmpeg -f concat -safe 0 -i concat.txt -c copy final.mp4

# upscale ONCE
quickdesign video upscale ./final.mp4 \
  --provider topaz --resolution 4k -o ./final-4k.mp4 --wait

Per-segment upscaling wastes ~Ncr × duration and can introduce inter-segment quality mismatches.

Gotchas

  1. Doesn't help with motion blur / shake — purely a resolution upscale. Pre-existing motion artifacts stay.
  2. Doesn't add detail — invents plausible textures from neighborhood context. Faces hold up reasonably; fine product detail (engraving, label text) can soften.
  3. 60fps inflation rarely worth the 2× cost for short-form social. Stay at native 24/30fps unless the platform requires otherwise.

Cross-references

  • UGC final concat step → ../pipelines/ugc-video.md (final concat section)
  • Subtitle post-processing (run BEFORE upscale or after, doesn't matter) → ../references/auto-subtitle.md

Source: SKILL.md on GitHub

2 warnings2mo3 checks · Risk SAFE
  • Gen Agent Trust Hub2mo

    This skill provides detailed instructions for using the `quickdesign` CLI to create and edit AI-generated media. It follows secure practices by implementing confirmation gates for paid tasks and provides guidance on interacting safely with downstream AI models. No security issues were detected.

  • Socket2mo

    1 alert: gptSecurity

  • Snyk2mo

    Risk: MEDIUM · 1 issue

Signed by skilld at 710a472. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last month.

Steadyupdated 3 months ago

README badge

README badge for anthropics/claude-plugins-community/quickdesign