Image to Video Prompts: Add Motion Without Warping
An image-to-video prompt should direct what changes over time. The starting image already supplies the subject, clothes and background. Upload it to CrePal image to video, then give one motion instruction and one preservation instruction. The examples below are prompts to test, not guaranteed outputs.
Describe the Motion, Not the Picture
Weak: “A woman wearing a blue coat in a beautiful city.” That repeats the still image. Stronger: “The woman turns her head toward the street as a bus passes behind her; camera remains fixed; her coat, face and buildings keep their shape.” The second instruction defines moving and stable elements. If the image contains small text or complex fingers, use less movement and inspect those areas in preview.
Camera Moves, Subject Motion and Mood Shifts: Prompt Building Blocks
| Motion type | Useful wording | Risk to watch |
|---|---|---|
| Camera | “Slow push in” | Edges stretch or background warps |
| Subject | “One natural blink” | Face changes between frames |
| Atmosphere | “Leaves move in a light breeze” | Static objects begin moving |
| Mood shift | “Clouds dim the light slightly” | Colors flicker or jump |
Start with a single motion type. “Slow orbit, zoom, pan and dramatic tilt” asks for incompatible camera behavior in a short clip.
Prompt Examples for a Portrait, a Product and a Landscape
Portrait: “The person gives one small smile and blinks; fixed camera; hair and clothing stay consistent; soft window light remains steady.”
Product: “Camera slides gently right across the ceramic mug; steam rises; the mug shape, handle and printed mark remain unchanged.”
Landscape: “A few clouds drift above the valley; camera makes a slow forward push; mountain outline and lake shoreline remain stable.”
For each, first use the sharpest available image with a clear subject. Preview the clip, note the first frame where something bends, and reduce motion or simplify the background on the next try. A 2×2 comparison of input and generated frames is more informative than a promise that “warping is fixed.”
Before/after prompt revision, shown as text rather than generated frames:
| Starting still | Overloaded motion request | Focused revision | Inspect in the resulting clip |
|---|---|---|---|
| Portrait by a window | “Turn, walk away, zoom and change the weather” | “One blink and slight head turn; fixed camera; background unchanged” | Face shape and window frame |
| Ceramic mug on a table | “Spin mug and fly around it” | “Slow camera slide to the right; mug stays fixed; steam rises” | Handle shape and printed mark |
| Valley landscape | “Storm arrives and camera orbits” | “Clouds drift slowly; gentle forward push; shoreline stays fixed” | Mountain line and water edge |
Use the focused revision as a new motion prompt. Keep the original still beside the exported clip and inspect three points:
| Point in the clip | Compare with the starting still | Revise if you see |
|---|---|---|
| First frame | Subject outline and background geometry | An immediate shape or crop jump |
| Middle frame | Face, product mark or shoreline | Identity drift or object repainting |
| Last frame | Lighting and intended motion endpoint | Flicker, unexpected movement or a cut that ends too early |
The prompt table compares instructions. Judge image quality only after comparing actual frames with the input still.
| What you see before or after generation | Change first | Why |
|---|---|---|
| The starting photo already has motion blur or a cropped subject | Repair or replace the still | The image itself gives conflicting or missing visual information |
| The still is clear but the subject bends midway through the clip | Reduce the requested subject and camera motion | Fewer simultaneous changes make the failure easier to isolate |
| A label or small text changes shape | Keep that area static or add exact lettering after export | Prompt wording alone may not preserve readable text |
Change the source when its visible cues contradict the intended action; change the motion prompt when a clear source image receives an overloaded instruction. This is a troubleshooting rule, not a guarantee of output quality.
From One Animated Image to a Full Multi-Scene Video
To make a sequence, choose a different authorized still for each scene and write a separate motion prompt per shot. Keep one visual anchor across shots: for example, the same matte green mug and pale oak table. This three-shot plan is an input example, not a generated gallery:
| Shot and starting still to prepare | Motion prompt to try | Compare before assembly |
|---|---|---|
| 1. Wide mug-and-table still | “Slow push toward the mug; table and handle stay fixed.” | Mug shape, handle and table edge |
| 2. Closer mug still from the same side | “Steam rises gently; camera remains still.” | Mug color and light direction against shot 1 |
| 3. Tight still of the same mug | “Very slight camera pullback; mug and surface remain unchanged.” | End framing and any unwanted text or repainting |
CrePal's image-to-video page describes uploading a starting frame, optionally an ending frame, describing motion, previewing and downloading or continuing into a larger project. Generate and review each clip separately; only then assemble the sequence and check the cut for continuity and timing.
Choose Motion the Source Image Can Support
Look at the still before writing the motion line. A tightly cropped portrait has little room for a sweeping orbit; ask for a small head turn or background movement. A product image with precise lettering may be safer with a fixed product and a slow camera push. A landscape can support cloud drift or moving water while the horizon stays level. In each case, name what must remain unchanged as explicitly as what should move.
For a portrait, try “subtle blink and a slight turn toward the window; camera fixed; face shape and clothing unchanged.” After the draft, inspect eyes, mouth, hairline and the first and last frames. If identity drifts, reduce subject motion or use a conventional pan over the original still. For a product, check label shape and edges frame by frame. A motion prompt is a test plan, not a guarantee that the generator will preserve every detail.
If the source frame is still being made, use the AI image prompt formula to settle its subject and crop first. For motion drift that persists after upload, use the prompt-to-video troubleshooting table to isolate the failed control.
CrePal Product Sources and How to Use Them
CrePal's image-to-video product page documents a starting frame, optional ending frame, motion description, preview and revision. Its official prompt guide recommends identifying subject action, camera move, pace and what stays stable. Use those fields as a review checklist after each draft; the guide's prompt examples are instructional text, not documented outputs for this page.
If the starting image shows a real person: review CrePal's Privacy Policy and Terms before upload, then keep the consent record with the source image. The motion table is about visual control; it does not establish permission to animate another person's likeness.
Related official reading: CrePal's image-to-video prompt guide. CrePal's dedicated image-to-video prompt guide uses subject action, one camera move, pace and a stability note as its compact pattern. For a portrait, keep motion small and facial features steady; for a product, review the label and final frame. Revise the failed motion choice rather than rewriting every instruction at once. Official prompt guidance, not a measured clip from this page.
Evidence and Review Notes
Evidence note added 5 November 2026. The prompts in this guide are editorial examples. The linked CrePal pages document product workflows or illustrative use cases; they do not verify an output from these exact prompts. No prompt-output pair is published here.
To document a result for this topic: Retain the licensed starting frame, motion prompt and clip; inspect first and last frames for object drift. Keep the exact prompt, model/version, settings, output URL, revisions and human review decision together.
Test your prompt in CrePal
Start with one example from this guide, then compare the result with your intended subject, style and details.
Site and support
CrePal's Terms identify SIGNOO PTE. LTD. as the service provider. The prompt examples here are instructions to try, not recorded generation tests.
Frequently Asked Questions
Large or conflicting motion requests can contribute to warped shapes; the model may also drift on complex details. Reduce to one movement, name the stable elements, and inspect the next result frame by frame.
Start with a clear, authorized image whose subject and background are easy to separate. Sharp edges and stable lighting make it easier to judge whether the animation preserves the original.
You can request a camera move such as “slow push in” or “gentle pan right.” Preview the clip, because prompt wording does not guarantee the exact path.