Vinyl and paper cutting
A shape a Cameo or Cricut can follow, without an evening spent tracing the outline by hand first.
One photo in · Clean cuttable edges · Nothing to install
One picture in, a solid black shape out. The outline stays exactly as the camera saw it — the stray hair, the gap under the arm, the tilt of the head — and everything inside it goes.
Made with the model this page uses. Drag the divider — watch the hair and the gap under the arm, which are the two places a bad conversion gives itself away.
Drag to compare before and after
Turning the brightness down produces a dark picture, not a silhouette. A real one is two tones: a solid black shape and pure white around it, with a boundary sharp enough to cut, print or trace. Everything the eye used to read inside the shape — face, clothing, texture — is gone, and what identifies the subject afterwards is entirely the outline.
This is one of those conversions people have always done by hand. Papercraft and Cameo owners trace outlines to feed a cutting machine, much as people do for a tattoo stencil, small businesses want a mark built from an actual photograph rather than stock clip art, teachers make name displays from the children's own profiles, and every year a run of people make a nursery print from the shape of a two-year-old's head. The tracing is the tedious part; the decision about which photo to trace was never difficult.
Backlighting in the original helps enormously and is not required. A subject in front of a window or a bright sky is already close to separated from its background, which leaves less for the model to decide. A photo taken against a busy room can still work — it just depends more on how clearly the subject's edge stands apart from what is behind it.
What you do with it next shapes what to ask for. For a cutting machine you want the fewest, cleanest closed curves possible, because every wobble becomes a blade movement. For a print you can afford fine detail like individual hairs. For a logo you want the shape simplified deliberately, since a mark that only reads at full size is not doing its job on a business card.
Objects and animals are usually easier than people. A bike, a tree, a cat mid-stretch, a bird on a wire — all have outlines that stay recognisable with nothing inside them, which is exactly the property a silhouette needs. A group of people standing shoulder to shoulder is the hardest case, because the moment their edges touch they merge into one shape.
Creative engine
Upload a picture, say what the shape is for, and the credit estimate updates before you press anything. The run happens on this page and the result opens directly under the box.
With the interior removed, a person facing the camera becomes an anonymous blob. Side-on, mid-stride, arms away from the body, a ponytail, a hat, a hand holding something — those are what make a silhouette readable, and they have to be in the photo before you upload it.
The space between an arm and a waist, between two legs, between the ear and a strand of hair. A conversion that fills those in has produced a shape that reads as a sack. Look at the negative space in the result before anything else.
Most of the quality is decided before you upload anything.
Turn side-on, get the arms away from the body, let the hair or the hat show. Once the interior is gone the outline is the only thing left carrying meaning.
A cutting machine wants few clean closed curves. A print can take individual strands of hair. A logo wants the shape deliberately simplified. Put it in the prompt.
Look at the gaps first — under the arm, between the legs, around the hair. If they are open and clean the shape will hold up at any size.
Mostly people who were going to trace it by hand tonight.
A shape a Cameo or Cricut can follow, without an evening spent tracing the outline by hand first.
A small business mark built from a real photograph — the actual building, the actual dog — instead of stock clip art everyone else also bought.
A child's own profile as a print, or a wall of thirty of them with names underneath. The traditional version of this involved a lamp and a pencil.
What survives, what to shoot, and what it is fit for afterwards.
Upload the picture on this page, say what you intend to do with the shape, and press generate. The result opens under the box as a solid black shape on white, ready to download. There is nothing to install and no tracing by hand.
It helps but it is not required. A subject in front of a window or a bright sky is already close to separated from its background, which leaves less to decide. A normally lit photo works too, as long as the subject's edge is clearly distinct from whatever is behind it.
Almost always the pose. Facing the camera straight on, with arms against the body, leaves an outline that could be anybody once the interior is removed. Turn side-on, move the arms away from the torso, and let something distinctive show — a ponytail, a hat, a hand holding something.
It should, and that is the first thing to check. Open negative space under an arm or between the legs is what separates a usable shape from a sack. If the gaps are filled in, run it again with a photo where they are more clearly visible.
Yes, and ask for it in the prompt — say you want the fewest, cleanest closed curves rather than fine detail, because every wobble in an outline becomes a blade movement. You may still need to convert the image to a vector in your cutting software.
Usually better than on people. A bike, a tree, a cat mid-stretch, a bird on a wire — all keep their meaning with nothing inside them. The hardest case is a group standing shoulder to shoulder, because touching edges merge into a single shape.
Yes, just ask for it. The same run can invert so the subject is white against a solid black field, which suits screen use and dark-background printing.
A standard image file you can download straight from the result, at the shape you chose before running. For cutting machines you will usually want to trace it to a vector in your own software afterwards; for print it is ready as it is.
Yes — take it into a [video model](/video/seedance-2) and the silhouette will animate, which is how the walking-figure clips people use as intros are usually made.
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. |
More conversions that remove rather than add, and the model behind them.
The outline you were going to spend an evening following, in about a minute. The estimate is on the button before you press it.
Examples made on Kavel with this model.
