dimeclips.

Dimeclips guide · Updated November 19, 2026

AI video prompts.
Give your scene
a payoff.

A giant kernel fills the cinema screen. Then the popcorn comes through it. One viewer quietly catches the last piece.

Give a short scene a clear opening, an action and an ending worth watching. Start with the exact prompt behind this campaign video, then use two simpler product examples to study motion and lighting.

Compare Dimeclips models and monthly allowances

“Front row was a mistake.” An original Dimeclips campaign scene.

AI-generated with Omni 1.2 directly on Google Cloud · 8-second source, 9-second finished edit · 720 × 1280 · Sound

01 / Build toward an ending

Set up the screen.
End with the catch.

Write the final visual beat before adding camera language. Here, the bucket catch makes sense of the chaos: the viewer gets exactly what they came for, in a very unexpected way.

  1. Opening · 0–1.5 secondsA giant kernel is already on screen. The first frame establishes the object that will change.
  2. Action · 1.5–6 secondsThe kernel pops and popcorn crosses into the audience. One impossible action drives the scene.
  3. Payoff · 6–8 secondsA calm viewer catches the final piece. Hold the reaction long enough to read it.

Exact generation prompt · Omni 1.2 Flash · 8 seconds · 9:16 · 720p

Original eight-second vertical cinematic comedy. A handheld phone camera behind two fictional adult cinema patrons looks toward an unbranded screen. The screen shows a giant popcorn kernel from the first frame. At 1.5 seconds it pops; by three seconds a fluffy wave of popcorn emerges physically through the screen toward the front row. Popcorn gently piles around the patrons without knocking them over. At six seconds one patron calmly raises an empty red bucket and catches one final piece. Hold the absurdly calm reaction. Maintain camera position, seats, people and screen geometry. Warm projector light, soft believable popcorn physics, original popping sounds and one surprised laugh. No dialogue, music, writing, logos or injury.

The opening caption, one-second dimeclips.com ending and audio level adjustment were added after generation. The timing above is the prompt’s plan; inspect how the actual video carries it out. Some popcorn briefly intersects clothing, a detail to catch during review.

This example was produced directly on Google Cloud for our campaign. The two examples below were generated through the Dimeclips product. Model, settings, input and editing choices all affect the result; repeating a prompt can produce a different take.

For your own scene, replace the kernel, screen and bucket with three connected elements: the setup, the surprising change and the reaction. Keep the camera and setting simple enough that the action stays readable. For ad concepts, use the ad workflow guide to record variants and review each attempt.

02 / Give the scene a direction

Five useful ingredients.

A prompt does not need to be long to be specific. Begin with a subject, then add the details that matter to the scene. You can use this structure as a starting point:

Subject + setting + motion + lighting + camera.

Subject
Name the thing the viewer should notice first, such as a yellow ceramic mug.
Setting
Place it somewhere concrete: a wooden table or a white pedestal.
Motion
Describe what changes over the clip. Rising steam is different from a turning mug.
Lighting
Describe the light you want to see, such as soft morning light or studio lighting.
Camera
Say whether the viewpoint stays still or moves. A static camera leaves the scene’s motion to the subject and light.

These are directions, not guaranteed controls. The model still interprets the scene, including details you leave unspecified.

03 / Product example: motion

The mug prompt.

This simple text-to-video example was generated in Dimeclips on November 16, 2026 with Wan 2.2 Fast, at a three-second setting and 832 × 480 resolution.

Exact prompt

A yellow ceramic mug on a wooden table, gentle steam rising, soft morning light, static camera, no text.

A yellow mug sits in a sunlit kitchen while steam rises above it. The camera stays in one position.

Actual Dimeclips output · Wan 2.2 Fast · 3 seconds · 832 × 480 · Silent

The mug gives the scene a clear subject. The table supplies a surface and context. “Gentle steam rising” asks for movement without asking the mug itself to travel, while “static camera” describes a steady viewpoint.

The result also includes a kitchen, a plant and a chair. Those details were not specified in the prompt. Look at that gap between the request and the result: it helps you decide what to describe more explicitly next time. If the background distracts you, a simpler setting is a reasonable next instruction to try; it is not a tested improvement in this example.

04 / Product example: light

The glass orb prompt.

Also generated in Dimeclips on November 16, 2026 with Wan 2.2 Fast, using the same three-second setting and 832 × 480 resolution.

Exact prompt

A blue glass orb on a white pedestal, gentle reflections moving across its surface, studio lighting, static camera, no text.

A blue glass orb rests on a white pedestal. Bright reflections and window-like shadows give the scene its visual texture.

Actual Dimeclips output · Wan 2.2 Fast · 3 seconds · 832 × 480 · Silent

Here, the requested movement is in the reflections. The prompt does not ask the orb to roll or the camera to orbit it. That distinction is useful when you want a quiet shot with something changing inside the frame.

The visible result adds strong window-like shadows and a warm background, even though the prompt only says “studio lighting.” This is another interpretation to inspect, not evidence that those details are predictable. Both examples end with “no text”; that is a request to the model, not a guarantee about every output.

05 / Make the next decision

Change one idea at a time.

After watching a result, identify the most important difference between your intention and what you see. Is the subject unclear? Is the background too busy? Is the movement different from what you wanted? Rewrite the relevant part of the prompt before adding more adjectives everywhere.

Keep the original prompt and result so you can compare them with your next attempt. Changing one instruction makes your comparison easier to understand, but a new generation can vary in other ways too. These two clips are examples of completed generations, not a controlled comparison or a promise that the same prompt will reproduce the same scene.

Plan for generation time as part of the workflow. Dimeclips generations run through a shared queue; a request may wait before processing. The short playback length of these examples does not represent how long they took to generate. Review each completed clip before deciding whether to keep it or try another version.