One photo in · A warm hug clip out · Free, no sign-up
Upload a photo of two people standing together and watch them turn toward each other and hug — the trend that turned old family photos into the most-shared clips on every feed.
Generated on this page's free engine from the still photo below the clip. Every output is AI-generated — your photo, your people and your hug will differ. Treat this as the shape of the result, not the exact frames.
From a photo of a grandmother and grandson standing side by side: they turn, she closes her eyes, he pats her back — six seconds that were never filmed but could have been. The room, the cardigan and the lamp light all carry over from the still.
Make your own
One frame, two people, both fully visible and facing the camera — the setup that gives the model its cleanest start. Scanned prints work exactly the same way; sharpness helps but sentiment does not require it.
Upload yoursIt is an image-to-video tool with one job: take a photo of two people and animate the moment they turn toward each other and embrace. The photo supplies the people, the place and the light; the model supplies six seconds of motion that never got filmed.
This search spiked for a reason that has nothing to do with technology: people found out it works on photos of relatives who are gone. A scanned print of a grandmother, a last vacation photo, a wedding-table snapshot — animated into one more hug. That is who is typing this, alongside couples in long distance and friends resurrecting an old graduation photo.
The emotional register is the whole product, which is why the pages that treat it as a generic 'video effect' miss.
The clip on this page is our own run on the free engine: a living-room photo of a grandmother and grandson standing side by side, and the sentence asked for a long warm hug with the setting kept. Watch what the model preserves — her cardigan pattern, the bookshelf, the lamp light — because that continuity is what separates a moving memory from an uncanny one.
The honest limit: hands and faces mid-embrace are the hardest frames in video generation, and some runs will show a blended arm or a softened face at the tightest moment. Rerunning deals a new take, and photos where both people are clearly separated and facing the camera give the model the cleanest start. A failed run refunds nothing because the free run cost nothing.
Creative engine
Upload the photo, keep or edit the pre-filled prompt, generate — the free engine returns about six seconds.
Room, clothes, lamp light — the clip keeps the photo's world and adds only the motion, which is why it reads as a memory.
A photographed print from the 80s animates the same way a phone photo does; sentiment does not require sharpness.
The photo choice does most of the work.
Two people, both faces visible, standing near each other. Old scans and phone photos both work — what matters is that the model can see who is hugging whom.
The pre-filled prompt asks for a long warm hug with the setting kept. Add the details that matter to you — "she closes her eyes", "he lifts her slightly" — the model listens.
Each run is a new take of the same embrace. Keep the one where the hands land right — with no account, the free credits cover the retakes.
Three photos, three very different rooms they get played in.
The clip of grandma hugging her grandson gets sent to the family group and watched more times than anyone admits. If the print is faded, restore the photo first — the hug lands harder in color. Bring a tissue.
The photo from the airport goodbye becomes the hug you are both counting down to — a better good-morning text than any emoji.
The graduation photo from years ago, animated into the reunion that has not happened yet. Send it with 'we're overdue' and watch plans finally form.
Asked before uploading the photo that matters.
Yes — this is quietly the most common use, and it deserves a straight answer. A clear scanned print animates the same way any photo does, and many people find the result genuinely comforting. Make it for yourself or your family; the one boundary is not making clips of people to mislead others, which our content rules and basic decency both cover.
One photo with both people in it works far better — the model animates what is already sharing a frame. If you only have separate photos, combine them side by side into one image first, then upload that; expect a less natural result than a true shared photo.
The free engine runs without an account: anonymous visitors hold free credits, and a hug clip costs a slice of them. Signing in adds a larger grant and keeps your clips in a history you can return to.
Around six seconds by default, selectable up to ten. Six is usually right — the turn, the embrace, a beat of holding — and shorter clips also render faster and spend fewer credits on retakes.
Uploaded over HTTPS to render the clip, stored so you can revisit it, never used to advertise. Delete it from your history and it is gone. Every output is AI-generated.
Interlocking limbs are the hardest thing a video model animates — two bodies briefly occupy the same space and fingers pay the price. It is the known weak point of this effect everywhere, not a settings problem. Rerun for a new take, and prefer photos where both people's arms are fully visible at the start.
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.
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 the supported image and video models (Nano Banana 2 / Nano Banana Pro / Seedance 2 / Veo 3.1), 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 2.0 | 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. Mini is cheaper still. |
| Video | Veo 3.1 | Billed per video (Lite). About 34 credits at 1080p, about 23 at 720p. |
| 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. About 8 credits per image. Start here to test. |
One photo, one sentence, six seconds that get saved forever — and if the moment calls for the romantic version, [the kiss generator](/video/ai-kiss-video-generator) is one door over.