
Upload a clear photo of yourself
Front-facing, both eyes visible, no sunglasses. This is the likeness everything else is built around, so it is worth using a good one.
One photo in · A believable snapshot out · Runs in the browser
The give-away in a fake is never the faces. It is that the two people are lit by different suns. This one matches the light first and the faces second.
Two generated snapshots. Look at the shadow direction across both faces rather than at the faces themselves — that is the part that decides whether a picture like this holds up. Every output is AI-generated and varies with your photo and prompt.


Drag to compare before and after
You upload one clear photo of your face. The model builds a new arm’s-length snapshot in which you are standing next to someone else, with a single consistent light source, phone-camera geometry, and your own likeness carried through unchanged.
Cutting one person out and pasting them next to another has been possible for thirty years, and it has never once been convincing. The reason is not the cutout — modern masking is excellent — it is that the two halves were photographed under different lights, at different focal lengths, at different distances. Your eye reads all three of those in a fraction of a second, even when you cannot name what is wrong.
Generating the whole frame at once removes the problem instead of patching it. There is no seam to hide because there are no two halves: the model renders both people into the same scene, under the same lamp, through the same lens. That is also why the prompt asks for the setting rather than for a background image — the setting decides the light, and the light decides whether it works.
The same machinery drives the plainer version of this on the AI selfie generator, where the second person is left out and you simply get a better shot of yourself. If you want to keep your own photo and only change how it was taken, the AI camera angle changer is the tool for that instead.
Creative engine
Upload your photo, name the second person and the setting, pick the format — the credit estimate updates before you generate.
Key light direction, hardness and colour temperature are solved once for the whole frame, which is the check your eye runs before it looks at anything else.
A wide lens held at arm's length gives a nearer shoulder, a slight barrel bend and a background that falls away — get it wrong and the picture reads as a poster.
The whole job is in step two, and it is one sentence long.

Front-facing, both eyes visible, no sunglasses. This is the likeness everything else is built around, so it is worth using a good one.

Who is standing next to you, and where. The place matters more than people expect — it sets the light for both of you.

One pass gives you the finished frame. If something feels off, it is almost always the light direction, and re-running with a different setting fixes it.
Mostly jokes, occasionally something with a real use.

The single most common use, by a wide margin: one friend, one improbable encounter, no explanation. It works because the picture is boring in exactly the way a real snapshot is boring.

Instead of a poster on the wall, a photograph in the camera roll. The appeal is the ordinariness of it — a corridor, bad light, someone half-blinking.
A historical figure, a character, a version of yourself from twenty years ago. Nothing about the tool requires the other person to be alive, or real.

Print it, frame it, hand it over. The joke lands harder on paper than it does in a message.

Generate one, show it to someone, and see how long it takes. Most people check the faces, which is the wrong place to look. If you want to practise on a simpler edit first, try the AI pose changer.

Drop the second person and keep the lighting work — the same pass produces a cleaner version of your own picture.
Mismatched light, nearly always. Two people photographed under different lamps and pasted together will never sit right, no matter how clean the cutout is. Generating the whole frame in one pass means there is only ever one lamp.
Good enough that your face is unambiguous: front-on, evenly lit, eyes visible, nothing across your face. Everything else in your photo gets discarded, so the background does not matter at all.
Yes, and you should. The setting is what sets the light — a bar at night and a street at noon produce completely different results from the same two people.
Yes. Describe them, or upload a second photo. Nothing in the tool is specific to well-known people; that is just the most common request.
Often, which is worth thinking about before you post one. Pictures like this are convincing precisely because they are unremarkable — plan for it to be taken at face value.
Credits are spent per generation, with the estimate shown before you commit. Paid credits open the stronger image models, which hold your likeness more reliably at small sizes.
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. Operator details, the full model list, and the refund window are on the about page.
Powered by
Use one balance across every supported image and video model — all of them 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. 20 credits per 1K image, 30 at 2K, 45 at 4K. |
| Image | Nano Banana Pro | Consistent run times across generations. 30 credits per 1K or 2K image, 50 at 4K. |
| Image | Nano Banana 2 Lite | Faster, simpler variant, and the everyday editing price: a flat 8 credits per image at every resolution. |
| Image | GPT Image 2 | The lowest-cost premium image model here. 10 credits per 1K image, 15 at 2K, 30 at 4K. |
| Image | Seedream 5.0 Lite | Flat pricing — 20 credits per image whether you generate or edit. |
One upload, one sentence about where you are standing. The shadows are the part to check when it comes back.