Creators chasing the format
Make a scroll-stopping reveal in minutes while the is-it-cake format is still moving, rather than baking, filming and editing the cut by hand.
One photo in · A short vertical clip out · No sign-up to try
A knife slices you open and there are sponge layers inside — the whole gag in one pass, your face intact right up to the cut, in a vertical clip ready to post.
These clips are not stock examples — we generated them on Kavel with the same one-photo image-to-video flow this page hands you, specifically to see whether the effect actually holds up. Every result is AI-generated, so yours will vary with the photo, the prompt, and the cake you set; treat this as the shape of the gag, not a template you get back exactly.
A plain full-body studio portrait went in and this came back — a knife slices down, the body opens into layered sponge and cream, and a slice lifts away on the blade while the face reads until the cut. One prompt, no editing, no second pass. It is also where the limits show: the effect is strongest on a clear, front-facing subject with the whole body in frame.
Open the video studioThe second run, to see how far the cake can be steered: a full-body portrait and a prompt asking for a chocolate layer cake with berries instead of plain sponge. The layers, the cream and the filling all landed, and the face carried over until the blade went in. Naming the cake is what changes the reveal rather than leaving it to chance.
Try your own cakeA knife presses in, slices down, and shows that the person was secretly a hyper-realistic cake the whole time — sponge layers and frosting where the body should be, usually in a few seconds and one clean cut. It animates that reveal from one still photo.
Nothing grim is happening: it is a playful, oddly-satisfying video gag, the moving-image version of the is-it-cake reveal, and it belongs next to the other meme transformations. On Kavel it runs image-to-video — you upload the picture, describe the cake you want, and the model keeps the face recognisable until the blade goes in, which is the entire reason the reveal lands.
We generated the clips on this page ourselves, on Kavel's own video studio and the same one-photo flow the page hands you: a plain full-body portrait went in, and a knife carved out a slice of layered sponge, in a single pass and a single prompt. That is the honest ceiling of the effect today — it is very good at a clean, front-facing subject against a plain background with the whole body in frame, and much less reliable when the cut has to work around props, busy scenes, or a tight crop.
Start from a sharp photo with the body fully in view; the flow defaults to image-to-video, and the credit estimate updates before you commit, so a failed run is refunded (unless it broke the content policy). It suits someone chasing the format while it is still moving, a creator who wants a scroll-stopping reveal, or a birthday post that turns the guest of honour into an actual cake.
Creative engine
Upload a photo, describe the cake, and the credit estimate updates before you generate.
Upload one clear photo and the model slices the body open to reveal sponge layers and frosting inside.
Describe the cake — vanilla sponge, chocolate layers, berry filling — and the clip follows that brief.
Check the credit estimate and generate; a failed run is refunded (unless it broke the content policy), so a second cake is free to try. Cakeify is the reveal version of the joke — the dance video generator keeps the subject whole and moving, and the kiss video generator is the gentle end of the same photo-to-video range.
Two things decide whether the reveal lands: what the blade has to cut through, and what it finds inside.
One clear picture, full body on a plain background — the blade needs the entire silhouette to travel through, and a cropped shot gives it nothing to open. A sharp front-facing photo keeps the face recognisable right up to the moment it stops being a face.
Vanilla sponge, a chocolate layer cake, a berry-filled cross-section — the interior is the punchline, so naming it is the difference between a clean reveal and an ambiguous smear. Say how the slice lifts away too.
The credit estimate shows before the run starts. Failed runs are free, so an interior that came back wrong costs a reworded brief and nothing else.
Trend accounts, birthday posts, anyone who saw the format and wanted a turn.
Make a scroll-stopping reveal in minutes while the is-it-cake format is still moving, rather than baking, filming and editing the cut by hand.
Turn the guest of honour into an actual cake and send the clip — the reveal lands because it is unmistakably their face right up until the knife goes in.
Common questions about the cake reveal clip on Kavel.
A video. The tool animates a still photo into a short vertical clip where a knife slices in and the reveal unfolds over a couple of seconds — the motion of the cut is the whole point, so a single frame would miss it. It comes back as a downloadable clip ready to post.
That is the reason for starting from a real photo — the model keeps the face and pose while the blade presses in, so it reads as you right up until the slice reveals the cake. A clear, well-lit original holds the likeness best; a dark or low-resolution one gives it less to work from.
A sharp, front-facing shot with the whole body in frame on a plain background — the cakeify cut needs the full figure to slice through, and a tight crop leaves it nowhere to go. Plain backdrops beat busy ones, and even studio light reads more cleanly than harsh shadow.
Yes. Describe it in the prompt — a vanilla sponge, a chocolate layer cake, a berry-filled cross-section, cream or ganache — and the reveal follows. A specific brief lands closer than a vague one, so it helps to name the layers and the filling rather than leaving them to chance.
Free credits arrive with the account, and the cakeify effect runs on Veo 3.1 — a free-tier model — so you can try it without a subscription. The estimate is on screen before you press the button, and a failed run is refunded automatically (unless it broke the content policy), so you can keep adjusting the brief until the reveal lands without watching a meter tick up on every attempt.
Props and busy scenes, for one — the model has to decide what is cake and what is not, and a cluttered frame confuses the cut. Tight crops, profile shots, and dark backgrounds all lower the hit rate too. If a run comes back mushy, the fix is almost always a cleaner, fuller-body source photo rather than a longer prompt.
One credit pool covers Nano Banana images and Seedance and Veo video. The cost shows before every run, and failed jobs are not charged. Use a subscription for ongoing work, or a one-time pack when you just need to top up.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
One-time top-ups — buy extra credits any time you run low.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
Charges appear as “KAVEL AI” on your card statement. You can cancel any time from Settings → Billing; cancellation takes effect at the end of the period you already paid for.
Powered by
Use one balance across every supported image and video model — eleven of them today, listed below — and check the credit cost before each request.
Video models
Image models
Model credit guide
You're only charged for successful generations. The exact estimate in the generator varies by model, length, resolution, audio, and number of images.
| Type | Model | Credit cost |
|---|---|---|
| Video | Seedance 2.5 | The long-take tier, priced per second: 5s 480p ≈ 211 credits, 5s 720p ≈ 473. A full 30s take runs ≈ 1,260 at 480p and ≈ 2,835 at 720p. Supplying a reference clip lowers the per-second rate. |
| Video | Seedance 2.0 | 5s 720p image-to-video ≈ 188 credits; text-to-video ≈ 308 credits. Scales with resolution and length. |
| Video | Seedance 2 Fast | Faster and lower cost. 5s 720p text-to-video ≈ 248 credits. |
| Video | Seedance 2 Mini | The cheapest Seedance tier. 5s 720p ≈ 154 credits, 5s 480p ≈ 72 credits. |
| Video | Seedance 1.5 Pro | Audio doubles the rate. 5s 720p ≈ 27 credits silent, ≈ 53 with audio; 1080p ≈ 57 and ≈ 113. |
| Video | Veo 3.1 | Billed per video, not per second (Lite tier). About 45 credits at 720p and 53 at 1080p. |
| Video | Kling 3.0 | Audio raises the rate. 5s 720p ≈ 105 credits silent, ≈ 150 with audio; 1080p ≈ 135 and ≈ 203. |
| Video | MiniMax H3 | Fixed 2K, no resolution ladder. Priced per second — a 5s clip ≈ 158 credits. |
| Image | Nano Banana 2 | Generate or edit from text and images. About 8 credits per 1K image, 12 at 2K, 18 at 4K. |
| Image | Nano Banana Pro | Consistent run times across generations. About 12 credits per 1K or 2K image, 21 at 4K. |
| Image | Nano Banana 2 Lite | Faster, simpler variant. A flat 8 credits per image at every resolution. Start here to test. |
| Image | GPT Image 2 | The lowest-cost image model here. About 3 credits per 1K image, 6 at 2K, 12 at 4K. |
| Image | Seedream 5.0 Lite | Flat pricing — about 8 credits per image whether you generate or edit. |
Other one-photo clips from the same video studio, and where to pick a model.
The whole joke is what the knife finds inside, so say what the cake is — then watch it open on a photo of you.
