If you can picture it, AI can make it.
Promptpins gives your AI the proven prompts to get there. See it, pin it, connect it.
Still poster → motion graphics in three directives (MiniMax H3)
A two-stage recipe: a strong still (here from @mmmiyama_D's poster prompt) + three plain-language motion directives, expanded to a full H3 prompt by an LLM. The insight: when the image is good, animating it is enough — don't redesign. Works for any typographic poster with a character.
@ponzponz15
SEO Rank Watch: an agent skill that improves one keyword a week until it ranks #1
A complete skill definition for a hands-off site: measure with Search Console → pick one near-#1 keyword → study intent → one improvement → 7-day cooldown → judge by data. The loop, the status states and the guardrails are what make it safe to run unattended. Requires your own GSC fetch script.
@yuno_miyako2
12-second picture-book commercial: Three Little Pigs → a straight corporate end card (MiniMax H3 Max)
A full spec for a 4:5 vertical 12s ad: timecoded shots, a 5-token palette (PAPER / INK / COOL / WARM / HOT), kinetic type, sound design and music. Swap {brand} and {headline} to reuse it for another business; the fable stays. Poster's cost: ¥24.
@ponzponz15
Photo × embroidery plate: a real photo on top, the same subject re-stitched below
Text-to-image, no input photo needed. Strong for single objects with clear outlines (chair, bird, lamp). The embroidery half tends to drift into felt or knit if the stitch names are removed — keep them.
@hann7712
Finger-frame video mask: show another clip inside the frame made by your hands
An agent instruction, not a generation prompt. Every line is geometry and order — no mood words — which is why it works. Ask for a 3-second check first.
@foxyy4i
Photo → minimal design poster (photo on top, geometric abstraction below)
Works when the subject has a recognizable silhouette. The bottom half extracts it into flat shapes; the palette is pulled from your photo. Weak for faces and busy scenes.
@didicvj7