1328 hand-tested AI prompts across image, video, and music — each with full context, variables, and a sample output. Content is authored in Traditional Chinese; the UI here is English.
三段式——第一段全是「攝影機規格」(the camera IS her eyes、完全不穩、…
第一行就用否定句 don't use one camera angle or single …
It first declares "locked, static camera," then uses a full sentence, "The camera never moves; every bit of motion is..." to pin down…
The sentence explicitly states "The audio leads the shot," which tells the model to build the visual rhythm around the sound…
Element counts are pinned down exactly (seven pairs of dot-work, two fern leaves), whitespace is written in metric units (at least 8mm between each element), and the finished size is also written in...
What makes a composite look fake always comes down to looking 'pasted on' — this line pins down, item by item, that it should compress over muscle bulges, blur slightly at the edges, and pick up the original photo's exposure and...
A complete skeleton for multi-shot consistency in Kling 3.0 — appearance is written once in the Master Prompt, individual shots...
Handles the two big requirements for wrapping paper and bag patterns at once: tileability, and staying unobtrusive as a background so a logo doesn't get lost in it. Achieved via low density plus even negative space plus low contrast to leave breathing room for a logo, and by banning directional lighting/shadow and perspective to preserve the tiling. Honest limitation: generative models don't guarantee true seamlessness — check the edge seams in image-editing software using an offset test, patch locally wherever there's a visible seam, and don't expect it to be perfect in one shot.
Generate a lavish key visual for a Ghost Festival (Zhongyuan Pudu) sale: a pyramid-style stack of offerings, warm candlelight, and promotional slogans and dates rendered accurately without garbled text — ready to use for hypermarkets, corner stores, group-buy organizers, and hang-tags.
Generate three OBS stream scenes in one go — starting-soon countdown, be-right-back, and stream-ending outro — sharing the same layout and color scheme with only the status text changing, with space reserved for the webcam and chat overlay, so your channel has one consistent brand look from stream start to sign-off.
Paste in a one-line project requirement, the deadline, and available headcount, and the AI produces a three-level WBS breakdown, a milestone timeline worked backward from the deadline, the critical path, and the top three risk alerts; if the deadline isn't feasible, it tells you straight whether to cut scope or add people.
Generate a professional turntable-quality 360° product video from text alone: constant angular rotation speed, a seamless loop with matching start and end frames, and continuous, non-flickering highlights on reflective materials — ready to use as the main video on a Shopee or momo listing page or as a livestream standby screen.
Generates a complete brand mascot character sheet in one shot with GPT Image 2 — the same character shown in multiple expressions and poses, side by side, in a consistent, reusable style, suited to small businesses making LINE stickers, customer-service avatars, or packaging identity.
Produces a complete character design sheet ready to hand off to an illustrator or use as a reference: front/side/back three-point views, a row of small expressions, accessory close-ups, and a Traditional Chinese spec panel — finalizing an original character from scratch.
A web admin dashboard mockup — left-side navigation, top-row KPI number cards, and a chart section — laid out on a grid so the information stays clear and uncrowded.
Break a user flow into 3-4 side-by-side phone screens with consistent components and style, so viewers instantly see how the flow plays out.
A three-column subscription pricing page mockup with plan cards whose prices, feature lists, and CTAs are all consistently aligned, and the middle column carrying a 'Most Popular' recommended badge.
Turn a packaging design into a 3D mockup: a box or bottle shown at a 45-degree angle revealing both front and side faces, with studio lighting at print-ready quality, and the brand name and product name rendered letter-for-letter on the package.
Exactly what architects, spatial planners, and early-stage real estate developers need — an exterior concept render to show owners or a jury before anything goes out to bid. Break the request into site context, building massing, facade materials, and time-of-day/weather, and GPT Image 2 produces a contextual exterior render suitable as a competition cover or fundraising deck hero image.
Exactly what short-film directors, ad scriptwriters, and branded micro-film pitches need — a single key frame that lets a client feel the tone of the film before it's shot. Break the request into shot type, lens, lighting, mood, and in-frame action, and GPT Image 2 produces a single storyboard frame in a cinematic 2.39:1 widescreen aspect ratio, ready to use as the key image in a pitch deck.
PRO21A young adult woman with soft refined features, thin metal glasses, and…
PRO22Turn this photo into a chaotic funny doodle illustration, intentionally…
PRO23Create an ultra-detailed 3D stylized creative designer poster featuring…
PRO24Use my uploaded face image as the primary identity reference. Preserve m…