Upload the photo · Ask for one change · Keep everything else
Almost every group photo has one person who blinked, looked away, or simply did not smile. This edits that one thing: the mouth softens into a real closed-lip smile and the eyes crease with it, while the skin texture, the hair, the clothes, the light and the background come back untouched. The test is not whether it smiles — it is whether anyone can tell it was edited.
Our own run on this page's model while building it. Drag the divider and look at what did not move: the stubble, the grey t-shirt's fabric texture, the black jeans, the concrete floor, the shadow on the wall. Every output is AI-generated, so your result varies with your photo and your prompt — read this as what the edit can hold.
Drag to compare before and after
It is a targeted expression edit: you upload a photo, ask for a smile, and the model regenerates the mouth and the area around the eyes while matching the original skin, lighting and detail so the rest of the picture is unchanged.
The old way of doing this was a warp — the corners of the mouth got dragged upward, which is why those edits always looked slightly rubbery and never touched the eyes. A generative edit rebuilds the region instead, so it can add the crease at the outer eye, the small change in the cheek, and the shadow under the lower lip that a real smile produces.
It also means the result is generated rather than moved, so a second run gives a slightly different smile.
Creative engine
Upload the photo, keep the preset prompt or name the smile you want, pick the shape. The credit estimate shows before you generate and a failed run is refunded automatically (unless it broke the content policy).
Closed-lip and warm, or open and showing teeth. Left unsaid, models default to a broad toothy grin, which is rarely what the photo needed.
Same face, hair, clothing, pose, lighting and background. Naming the constants is what keeps the edit local — this is the single biggest difference between a believable result and a new person.
Generated teeth are where these edits give themselves away. If a toothy smile comes back uneven, ask for a closed-lip smile instead — it is both easier and usually more natural in a portrait.
For a smile in motion rather than a still, the laughing face effect animates the whole reaction; if the teeth themselves are the point, the braces filter works on them directly.
Three things to write down. Leave them out and you get a different, happier stranger.
Change only the expression. Anything broader — make him look happier, improve the photo — invites the model to re-light the face or move the head, and then the edit is obvious.
A natural, warm closed-lip smile with a slight crinkle at the eyes. That sentence is doing three jobs: setting the type, keeping the teeth out of it, and asking for the eye movement that makes it read as real.
Keep the exact same face, skin texture, hair, clothing, pose, lighting and background. It feels redundant to write and it is the reason the after frame still looks like the same photograph.
Three cases where reshooting is not an option.
Everybody is fine except one person, and there is no second attempt. Fixing one expression is far less invasive than compositing a face from another frame.
A stiff expression is the most common reason a good headshot gets rejected. A closed-lip smile is the safe register here, and it keeps the shot usable on a profile.
Formal portraits from an era when nobody smiled at cameras. Restore the photo first with old photo restoration, then decide whether the expression should change at all.
What it keeps, where it fails, what a run costs.
It does when the prompt pins the constants — same face, skin texture, hair, clothing, lighting, background. Our own run kept the stubble, the fabric texture of the t-shirt and the shadow on the wall unchanged; the only difference between the two frames on this page is the mouth and eyes.
Yes, and it is the riskier option. Generated teeth are the most common tell in this kind of edit — uneven, too white, or one tooth too many. Ask for it if the photo needs it, and be ready to fall back to a closed-lip smile.
It works, but be specific about who — the person on the left, in the blue shirt. Without that the model may adjust more than one face, and then the edit stops being invisible.
Less well. The edit has to match the surrounding detail, so if there is no detail to match, the regenerated area often comes back sharper than the rest of the face — which reads as fake. Upscale first, then edit.
That is your call, and worth making deliberately. Fixing your own family photo is not the same as altering a picture of someone else for something they did not consent to. We would not use it on a photo intended as a record of an event.
Free credits come with the account and cover the first attempts. After that each edit costs credits like any image, the estimate for your model and size appears in the generator before you start, and a failed run is refunded automatically (unless it broke the content policy).
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. |
Three tools that change a single element and leave the photograph alone.
Upload the photo, ask for one change, keep everything else. Free credits cover the first tries.
A real edit made with this tool — drag your own photo in on the left, the prompt is already written, then press Generate.
