Artistic Ink Landscape Transformation
A detailed prompt for generating a high-quality video where falling ink splashes evolve into a vivid miniature landscape painting.
87 prompts
A detailed prompt for generating a high-quality video where falling ink splashes evolve into a vivid miniature landscape painting.
A creative video prompt for generating bootleg-style footage of a shark rock band performing on surfboards in a dive bar setting.
A highly detailed anime-style short video prompt featuring a rooftop soccer scene with precise timestamped actions and audio guidance.
A comprehensive multi-scene prompt for a hyper-realistic ASMR video of a LEGO Lamborghini assembly, including camera and sound instructions.
A specific prompt for Gemini Omni Flash designed to replace video objects with a cat from a reference image while maintaining scene consistency.
A creative prompt for making a video similar to The Sims character creation interface, featuring dancing, a rotating Plumbob, and static UI.
A technical prompt for generating a fast-paced 10-second video featuring high-energy match cuts at a rate of two per second in a cinematic 16:9 aspect ratio.
A high-end fashion commercial prompt featuring toddlers in a luxury aesthetic described as feral and unstoppable.
A creative video generation prompt depicting black-and-white ink wash shrimp leaping from rice paper into a futuristic neon city with glowing orbital tentacles.
A detailed image-to-video prompt where a woman breathes fire at a birthday party, ending with a black-and-white slow-motion effect.
A comprehensive JSON prompt for creating a high-speed cinematic drone commercial featuring honey packaging set against the Himalayas.
A descriptive cinematic prompt for generating a 10-second time-lapse of an indoor living room renovation, featuring light changes and realistic construction.
A prompt depicting a traditional Japanese sumo wrestling scene with rikishi in mawashi and chonmage hairstyles, focusing on the pre-fight ritual.
A high-detail motion graphics prompt for Gemini Omni to create a celebratory animation for X earnings milestones, featuring 3D rotations, particles, and synchronized typography.
A detailed multi-scene prompt for video generation models featuring an orange American Shorthair cat acting as a professional barista in a coffee shop.
A highly detailed prompt for a high-end summer beverage commercial featuring macro product close-ups, golden hour lighting, and poolside lifestyle scenes.
A cinematic video prompt designed for Gemini to generate high-end time-lapse footage of Indonesian chicken noodles (Mie Ayam) with professional lighting and macro details.
A video-to-video editing prompt for Gemini Omni designed to modify scenes to include Chanel products and elegant hotel suites while maintaining the subject's appearance.
A highly detailed cinematic video prompt for Gemini Omni featuring a mechanical mantis made from a pepper grinder fighting a cockroach in a kitchen.
A whimsical Gemini Omni prompt for a video of a cat in a bubble helmet surrounded by fish, creating a realistic yet dreamlike scene.
A high-quality video prompt for a fitness soda commercial using the AIDA model to describe post-workout recovery with macro shots and electric blue effects.
A comprehensive video generation prompt for Gemini Omni featuring detailed school drama scenes with specific camera movements and dialogue.
A creative video generation prompt where a soccer player kicks a ball out of the screen and then jumps out to perform kick-ups on a person's palm.
A detailed cinematic prompt for Gemini Omni to create a 10-second suspenseful video of finding a baby penguin in a treasure chest.
Last reviewed August 17, 2026 · Editorial synthesis of current Topview samples and public prompting guidance
Quick answer
A strong Gemini Omni prompt starts with the creative goal, then turns that goal into visible action, camera behavior, and trackable sound cues. For a simple shot, identify the subject, the change that happens, the framing, the dominant visual style, and the ending. For a multi-beat concept, add short time ranges or an ordered storyboard so every beat has one job. When the selected Topview mode accepts reference inputs, label each input by role—identity, product shape, environment, motion, style, or rhythm—and state which details should be preserved and which may change. Avoid making every reference control everything. Audio direction is clearer when dialogue, ambience, music, and effects are separated and important cues are tied to visible events. Editing prompts should be narrower than generation prompts: name the one or two changes, then restate the camera, subject, timing, or layout that should remain. The Topview sample library ranges from concise single-shot briefs to detailed timelines, product commercials, storyboard-driven clips, and object-replacement edits, so prompt length should follow the number of production decisions you need to control. Generate a draft, review action order, framing, input use, audio timing, readable text, and the last frame, then revise the smallest unclear instruction first.
[goal and format] + [input roles when available] + [subject and action] + [timeline or storyboard] + [camera plan] + [audio and sync cues] + [style and lighting] + [preserve and change rules]
Name the intended result, audience or channel, duration, and aspect ratio before describing individual details.
When the selected Topview mode accepts reference inputs, label each image, video, audio file, or storyboard and give it one primary responsibility.
Describe who or what the viewer follows, what changes on screen, and the visible cause-and-effect that carries the shot.
Use ordered beats for a sequence; reserve timestamps for moments where order, pacing, or synchronization is important.
Choose framing, point of view, and one motivated camera behavior per shot, such as locked-off, push-in, orbit, tracking, or handheld follow.
Separate dialogue, ambience, music, and effects, then connect key sounds to an action, edit, word, or time marker.
Use one dominant visual system and support it with a coherent palette, material language, lens feel, and light source.
For revisions or reference-led work, list the few elements allowed to change and the details you want the result to retain.
Match the amount of structure to the idea. A single visual event may need one shot brief; a narrative or commercial usually benefits from ordered beats.
| Goal | Prompt pattern | Example wording |
|---|---|---|
| Direct one continuous action | Name the starting state, one visible trigger, the resulting change, and the state held at the end. | One continuous 8-second macro shot: a drop touches the paper flower, its folded petals open in sequence, then the completed bloom holds for the final second. |
| Build a short narrative arc | Divide the duration into setup, development, payoff, and final hold; give each range one readable action. | 0–2s: flashlight finds a mossy chest. 2–6s: a hand opens it. 6–9s: a tiny penguin looks up and waves. 9–10s: close-up hold. |
| Create a product reveal | Move from context to material detail to a stable hero frame while protecting the product description. | Begin with a fast aerial approach, slow into a controlled 90-degree orbit, then end on a centered macro hero shot with the package front readable. |
| Explain a concept | State the idea, the audience level, the visual metaphor, and the exact labels or facts that matter. | Explain refraction for a beginner using a glass prism and one white beam; show the beam separating into a labeled spectrum in a clean editorial animation. |
| Animate on-screen text | Provide exact words, order, placement, exposure time, motion style, and the rhythm that controls each appearance. | Show one centered word per beat—“MAKE”, “THE”, “IDEA”, “MOVE”—white type on black, each entering with a different geometric wipe, no other text. |
When the selected Topview mode accepts the relevant reference type, assign every input a bounded job. Availability can differ by mode, so use only the inputs shown in the current interface.
| Goal | Prompt pattern | Example wording |
|---|---|---|
| Use an identity reference | When the selected Topview mode accepts image references, say that the image defines the person and list the identity details you want to preserve. | [Image 1] defines the protagonist: facial features, short silver hair, skin tone, charcoal coat, and amber ear cuff; use the prompt for action and setting. |
| Protect product design | When the selected Topview mode accepts product images, give one image responsibility for shape, material, color, and label placement. | [Image 2] controls the bottle proportions, coral glass, black cap, and front-label position; lighting and background may change. |
| Borrow motion from video | When the selected Topview mode accepts video references, specify the motion or camera path to borrow and the appearance not to transfer. | [Video 1] supplies the dancer’s timing and camera path only; use the subject and wardrobe from [Image 1], not the people or colors from the video. |
| Follow an audio rhythm | When the selected Topview mode accepts audio references, identify whether the file controls tempo, edit points, delivery, or mood. | [Audio 1] controls rhythm only; place the four scene changes on its four strongest percussion accents and keep dialogue absent. |
Camera and sound directions work best when they support the same story beat. Use named, observable cues instead of broad terms such as “dynamic” or “epic.”
| Goal | Prompt pattern | Example wording |
|---|---|---|
| Keep attention on one action | Use a locked-off frame or a restrained push-in and let subject motion provide the change. | Locked medium shot; the ceramic bird unfolds its wings as the camera remains still, with soft hinge clicks in sync with each fold. |
| Reveal scale | Begin on a close detail, then widen or pull back after the trigger that makes the larger setting meaningful. | Macro close-up of frost on a window; when the first train horn sounds, pull back to reveal the entire snowbound station. |
| Create grounded energy | Describe the operator behavior, subject direction, and realistic sound bed instead of stacking several camera moves. | Shoulder-height handheld follow as the cyclist travels left to right; subtle footstep sway, tire hiss, chain clicks, and mountain wind. |
| Synchronize a transformation | Name the visible state change and attach its start, peak, or completion to a distinct sound. | At 5s, the paper skyline folds into a real city on the bass hit; add one paper snap at the start and let traffic ambience fade in afterward. |
| Direct dialogue | Identify the speaker, exact words, performance, framing, and background sound; keep the line short enough for the shot. |
When the selected Topview mode accepts a storyboard or existing media, use it as a map, then state the reading order, timing, allowed changes, and protected details.
| Goal | Prompt pattern | Example wording |
|---|---|---|
| Turn a storyboard into video | When the selected Topview mode accepts storyboard images, state panel order, total duration, shot timing, and whether transitions should be cuts or continuous moves. | Follow panels 1–6 from top left to bottom right within 10 seconds; preserve the character design and panel compositions, using clean cuts between beats. |
| Replace one object | When the selected Topview mode accepts an input video, name the replacement and protect motion, timing, camera, and environment. | Change only the red umbrella to a translucent paper parasol; keep the walker, pace, rain, camera path, street layout, and clip timing unchanged. |
| Restyle existing motion | When the selected Topview mode accepts an input video, separate the visual treatment to change from the action and camera to retain. | Apply charcoal-and-watercolor textures to the existing clip; retain the original walk cycle, cut points, framing, and left-to-right screen direction. |
| Change the camera treatment | When the selected Topview mode accepts an input video, request one framing or movement change and restate the subject action that should remain. | Keep the violin performance and timing; change the view to an over-the-shoulder medium shot with one slow push-in, no additional cuts. |
Make an amazing cinematic product video using my references. Add cool camera moves, dramatic transformations, perfect music, clear text, and keep everything consistent.
10-second 9:16 launch teaser for a coral-glass fragrance bottle. When the selected Topview mode accepts image references, [Image 1] defines the bottle shape, black cap, front-label placement, and coral glass; keep those product details recognizable while the set and lighting change. [0–2s] Locked macro shot of condensation forming on the bottle, quiet room tone. [2–5s] One water drop reaches the base and the white tabletop unfolds into layered paper waves; slow pull-back, one crisp paper-fold effect. [5–8s] The waves settle into a dark blue studio as the camera completes a controlled 90-degree orbit. When the selected Topview mode accepts audio references, use [Audio 1] for beat timing only and land the environment change on its strongest accent; otherwise use a written low electronic pulse at 5s. [8–10s] Centered hero frame, front label facing camera, soft cyan rim light, exact on-screen text “STILL / MOVING” in two lines for the final second. Preserve product geometry and the coral-blue palette. No extra products, people, captions, or additional logos.
The weak prompt does not define what any reference controls, what the product is, how the transformation develops, which camera move matters, when sound should land, what “consistent” means, or how the clip ends. The stronger prompt assigns bounded reference roles conditionally, gives each time range one action, coordinates camera and sound around the transformation, defines exact text, and separates protected product details from the environment that may change. These instructions create specific review points without assuming that a generation will follow every detail perfectly.
Name the deliverable, subject, and visible change first; add style only after the action is clear.
When the selected Topview mode accepts references, assign each input one primary role and resolve conflicts in writing.
Add a starting state, one observable action or trigger, and a final state the viewer can recognize.
Rank the story beats, remove any beat that does not support the payoff, and give the remaining actions enough time to read.
Choose a shot size, point of view, and one motivated movement for each beat.
Separate the sound layers and tie the most important cue to a word, action, cut, or timestamp.
Change one or two variables per revision and restate the camera, subject, timing, and layout you want to retain.
Describe the specific features to preserve, keep the action manageable, and review those features after every generation.
Define the final framing, subject state, motion state, sound state, and any exact text that should remain visible.
A useful prompt states the goal, subject, visible action, camera behavior, sound intent, visual treatment, and ending. Add input roles, timestamps, or preserve rules only when they remove a real production ambiguity.
Use the shortest prompt that still resolves the important decisions. A single continuous shot can be compact; a multi-shot commercial, storyboard, synchronized audio sequence, or controlled edit usually needs more structure.
When the selected Topview mode accepts those input types, label each reference and specify whether it controls identity, product shape, environment, motion, style, or rhythm. Do not assume every mode accepts every input type, and do not ask one reference to control conflicting layers.
When the selected Topview mode accepts storyboard input, state the panel reading order, total duration, shot timing, transition logic, and the character or product details to retain. If no storyboard input is available, translate the panels into numbered written beats.
When the selected Topview mode accepts existing media for editing, name one or two changes and then list what should remain: subject, motion, timing, camera, background layout, lighting direction, or audio. Narrow follow-ups are easier to compare than a full rewrite.
Tie the most important visible event to a distinct beat, word, sound effect, or timestamp. When the selected Topview mode accepts audio references, state whether the audio controls tempo, cut points, delivery, or mood; otherwise describe the intended timing in text.
Use one stable description, identify the exact features to preserve, avoid unnecessary scene and wardrobe changes, and repeat the preserve list after major cuts. When the selected Topview mode accepts reference images, give a clean identity or product image a single role, then review the output rather than treating the reference as a guarantee.
The prompt may contain more actions than the duration can show, conflicting input roles, incompatible styles, or several camera and sound directions competing in one beat. Remove lower-priority instructions, clarify one change at a time, and retry with the successful parts restated.
How this guide was built
This guide is an original editorial synthesis of recurring structures in Topview’s public Gemini Omni sample library—single-shot action briefs, timed narratives, camera-led product reveals, sound-aware scenes, storyboard workflows, reference-led continuity requests, and targeted video edits—reviewed on August 17, 2026. It was cross-checked against Google DeepMind’s public Gemini Omni prompting guidance on framing and motion, action, style, lighting, conversational editing, camera direction, multimodal references, audio synchronization, text, and storyboards, then expanded to cover the search intent commonly addressed by public prompt-guide pages: formula, examples, weak-versus-strong comparison, mistakes, workflow, and FAQ. The examples are writing patterns, not benchmark claims. Input support, limits, model behavior, and output quality can vary by selected Topview mode and may change over time.
| Combine multiple inputs |
|---|
| When the selected Topview mode accepts multiple inputs, map each one to a different layer and resolve any conflict explicitly. |
| [Image 1] defines identity, [Image 2] defines the rooftop layout, and [Audio 1] defines beat timing; the written prompt controls action, camera, and the blue-gold palette. |
| Static close-up. The shopkeeper says quietly, “We close when the rain stops.” No subtitles; rain and refrigerator hum remain below the voice. |
| Make an iterative revision | Change one or two layers per pass and list the previously successful layers to preserve for comparison. | Keep the subject, action order, camera path, bottle geometry, and label placement; change only the room to a dark mirrored studio and add a cool rim light. |
|---|