Try it with the README you already have:
“Make a 30-second product intro from this README.”
Install + original sample prompts:
github.com/JakeB-5/motion-gr…
Play the demo, no install:
jakeb-5.github.io/motion-gra…
Same skill. Three very different briefs.
A product launch, a space-telescope explainer, a coffee guide.
These are frames from the actual outputs. Each sample includes its original prompt, storyboard and source—so you can try it with your own material.
Your README could be your next product video.
I built a Codex + Claude Code skill that turns source material into motion graphics—with music, in one HTML file.
This 30s demo was made with it. Sound on 🔊
Open source. Prompt + examples below ↓
Just realized the math on that 5% context tax on Max 20x:
• Max 5x: 20% gone instantly
• Pro (1x): 100% wiped out.
If switch accounts on Pro and Claude literally says "Congratulations, you played yourself. See you in 5 hours." 🤡
Switched accounts mid-session, and Claude immediately ate 5% of my Max quota just to re-ingest the exact same context. Guess namespacing user IDs into prompt cache keys is more important than actual UX. Outstanding move. 👏
#Claude #Anthropic
I already had my entire workflow aggressively optimized for caching, and with the 75% price drop, the billing stats are undeniably lower. So why the hell has my weekly limit already burned past 50%? fable 5.1
@로 파일을 입력하는건 컨텍스트에 파일전체내용을 강제로 주입하는것인데, 파일 이름으로 전달하는것에 비해서 이득이 될 부분이 있나요? 일부 소형파일에서는 툴콜하는것에 비해 이득이 있어보이지만 대부분의 경우 비효율같아서요
Bragging about a 75% price drop with Fable 5.1 is hilarious when the output token speed has crawled to a complete halt. Who gives a damn about 75% cheaper cache reads if we have to grow old waiting for a single response to finish generating?