Guides
Turn Screenshots and Scripts Into AI Motion Graphics
A screenshot supplies visible evidence; a script supplies purpose and sequence. Neither needs to describe every animation detail, but both should identify what must remain accurate. In Motion AI Video, start with text and optionally add up to four reference images. Use the two hypothetical briefs below as working patterns, replacing their product details and on-screen wording with your own approved material.
Separate fixed content from creative choices
List the names, interface labels, numbers, quotations and brand elements that must stay unchanged. Then describe the creative freedom: headline placement, surrounding space, entrance style or emphasis. A screenshot can be drawn as the original image with effects over it; do not assume automatic cutouts or individually editable controls. Asking for preservation gives you a review target, not a guarantee that inspection is unnecessary.
Prepare one useful source view
Use PNG, JPEG or WebP files, each no larger than 5 MiB. Crop or redact in your image editor before uploading. Keep enough context to explain the interface, while making the important panel readable at the destination size. For multiple images, name each image’s role in the prompt instead of expecting the system to infer their order.
- Replace private data with approved demonstration data and remove notifications.
- Check the actual product state, all required labels and the logo before generating.
- Supply standalone brand artwork as a supported raster image; convert SVG, PDF or slide material beforehand.
Prompt example: a product screenshot
Select 15 seconds and 16:9 before submitting this hypothetical reporting-tool brief. Its panel name must match your actual screenshot.
Create a 15-second product video for a team lead. Use the supplied dashboard screenshot as the main image. Open with the exact text “A clearer view of the work.” Introduce the screenshot, then gently emphasize the panel labeled “Monthly overview.” Preserve the original labels, values, logo and visible controls. Keep the screenshot large and give viewers a quiet reading interval. Finish with “See the work in context.” Use restrained motion and space around the headline. Do not invent metrics, controls or features.
Turn a script into an ordered brief
Paste the script as text instructions. Separate exact on-screen copy from narration and visual directions, and put each direction beside its relevant moment. Remove competing ideas before generating. Built-in music, uploaded audio and AI voiceover are available; narration should support a sequence that still makes sense without sound. AI voiceover uses subscription credits separately.
Prompt example: a text-led explainer
Choose 30 seconds for this hypothetical employee onboarding sequence. Select the destination’s aspect ratio before generation; it cannot be changed later through chat.
Create a 30-second explainer for a new employee. Open with “A clear path for every request.” Show these exact stages in order: “Submit,” with “Share the details”; “Review,” with “Confirm the next action”; and “Resolve,” with “Close the loop.” Give each stage its own moment before showing the connected sequence. Use clean typography, consistent cards and readable supporting phrases. Finish with “Know what happens next.” Do not add stages, response-time promises or decorative motion competing with the labels.
Review accuracy, then refine the motion
Compare wording, numbers and screenshot details with the approved source first. Then watch at normal speed and at the intended viewing size. Ask for one concrete revision, such as a quieter panel highlight or a longer reading hold within the selected duration. Compare the new version with history and recheck preserved details. Export a 1080p MP4 in a supported browser and review the actual file before publishing.
Write a targeted revision request