How to create
scroll-stopping
AI social videos.

Download Credits are for hand-crafted animations made by our in-house motion crew. AI Credits are for AI videos. Totally separate, never double-charged.

We always show you exactly how many credits you're using, right above the Animate button.

You get multiple videos at no extra credit cost:

• Pick 2 videos - same credits as only 1 video
• Pick 4 videos - same credits as only 2 videos

Free regeneration: Every video includes 1 free regeneration.

The initial view is a static scene preview - confirm the subject, caption, and overall visual setup before animating.

There's no way to preview motion beforehand. The animation step is where the heavy AI rendering happens, which is why credits are consumed here - rather than at download. This is standard across all AI video platforms (Runway, Pika, Higgsfield, Google Veo, Luma Dream Machine all work the same way).

Make the most of your scene preview: Use it as a checkpoint to ensure everything looks perfect. Scene previews don't use any credits - edit your prompt or regenerate until you're happy. Before you animate, check these points.

Yes. Upload logos, products, photos, or designs at the Prompt step. Important: you must tell the AI how to use it in your prompt. Say "my uploaded image" and specify where/how it should be integrated into the scene.

Example prompts:

Scene to avoid · No clear subject

No subject that can pop out.

The pop-out effect needs something that can physically move toward the viewer. An empty room, a building exterior, or a fixed object like a sign bolted to a wall or a clamp locked to a wheel - these can't pop out in real life, so they can't pop out in your video either.

Fix: Add a subject to your scene that can realistically pop out - a person, animal, character, or movable object.

Scene to avoid · Barrier blocking the path

Something is between the subject and the camera.

If there's a desk, counter, table, or any object between your subject and the camera, they can't pop through it. The subject needs a clear path forward. Think of it this way: the AI follows real-world physics - if a barrier would stop someone walking toward you in real life, it will stop them in your video too.

Fix: If the AI keeps adding a barrier, be explicit in your prompt: "no desk", "standing in open space", or "nothing between subject and camera".

Scene to avoid · Awkward angle

Product held at an awkward angle.

Look at the woman holding the product below - how would that product naturally pop out? If she brought it forward, her arm would need to pass directly through the white social post area below instead of over it, which doesn't make for a good pop-out. The framing is too tight. You need to zoom out and allow more space, giving her arm the room it needs to extend naturally toward the camera.

Fix: Position the product at chest or waist level, with enough spacing in the scene for everything to naturally pop out. Add the below phrases to your prompt to force the right framing.

Close up shot

Add close up shot for hip-to-head framing (more space to allow a clean pop-out). Best for product shots.

Full body shot

Add full body shot. Best when holding a big enough product - small products held far back won't be clear when they pop out.

Scene to avoid · Framing too close

Camera is too close to the subject.

If the camera is so close that only the face fills the frame, what exactly would pop out? A disembodied head flying toward the viewer isn't the effect you're after. The pop-out works best when there is adequate spacing - you can see the subject's upper body (at minimum their arms, torso and head) so there's something substantial to pop out.

Fix: Pull back the camera to allow more space in the scene. Add the below phrases to your prompt to force the right framing.

Close up shot

Add close up shot to get hip-to-head framing.

Full body shot

Add full body shot to get full body framing.

The rule: If you can't immediately answer both questions when looking at your scene preview, regenerate until you can. Scene previews are free - use them.

That branding lives in the original prompt. Go back to the Prompt step, find the text in quotation marks or the uploaded logo, replace it with your own, then regenerate. Learn more.

This happens when text isn't clearly readable in the scene preview presented to you prior to animating.

If you want text or a logo to come through with clarity, it needs to be big, bold, visible, and not obstructed by anything in your scene preview. Scene previews don't use any credits - edit your prompt or regenerate until everything is clearly visible.

The fix: If you can't easily read the text in your preview, try adding "close-up shot" to your prompt to bring it closer. Or, write your prompt around the focus point, emphasising that it needs to be big, bold and visible, like in this example for a book cover:

Same rule as text - if a logo isn't big, bold, and clearly readable in the scene preview, it won't render cleanly in the final video.

The edge case: logos with a high text density. If your logo has small text - a tagline beneath the brand name, a slogan, or just a long brand name that has to shrink to fit - it often becomes too tiny to render clearly once placed in a scene. The AI is then forced to guess at that text, and as a result it may look distorted.

Example prompt using the brand name in quotes instead of uploading the full logo:

We have features baked into the experience to solve most issues, and they're included with every video.

You get more videos, without using any extra credits:

• Pick 2 videos - uses same credits as only 1 video
• Pick 4 videos - uses same credits as only 2 videos

All videos render at the same time. Compare variations, pick your favourite, ignore the rest. The extras didn't cost you anything.

AI video is creative and unpredictable - sometimes it nails it, sometimes it doesn't. That's why we give you multiple videos at no extra credit cost + free regens upfront instead of charging per try.

If every video has the same issue, it's usually a prompting problem.

Submit a ticket to [email protected] with:

• Your exact prompt, and the video name in your account
• What is wrong with the video
• Whether this happened in all variations or just one

One of our friendly team members will assist you further :)

Every AI-generated video is unique - even with the exact same prompt, you'll never get a 100% identical result twice. This is how all AI video platforms work (Runway, Higgsfield, Google Veo, Pika, Luma included) - submit the same prompt twice on any platform and you'll get two different videos. Think of it like asking an artist to paint the same scene twice - the overall composition will be similar, but the details will always vary.

When you remix, you're reusing the same prompt and scene setup as a starting point - you're not copying the final animation. The text, images, and motion in an AI video are all generated together in a single pass rather than as separate layers, so there's no way to freeze the animation and just swap one element. Every render produces a unique result.

Use multi-generation to get several variations and pick the one closest to what you're after.

Need exact consistency? Use our hand-crafted animations made by our in-house motion crew - 2,000+ templates with predictable results.

Unfortunately not - with AI video generation, the text, images, and motion are all generated together in a single pass rather than as separate layers. There's no way to freeze the animation and just swap one element - any edit means a fresh render with a new, unique animation. This is how all AI video platforms work (HeyGen, Google Veo, Higgsfield and others all work the same way).

To change text, it depends on where the text is:

• Text inside the scene (signs, billboards, t-shirts) - Go back to the Prompt step, find the text, replace it, and resubmit the prompt.
• Text in the social post caption (username, caption, reactions) - Edit at the Layout step.

Every change requires a re-render. This is why you get a scene preview first - so you can double-check everything before using credits.

Yes, it's available in the Prompt step. There's a pill button directly below the prompt box where you can select your preferred aspect ratio.