Last frame in · The next seconds out · Free engine by default
Continue a clip past the frame where it stopped — upload the last frame, describe what happens next, and the scene keeps going instead of starting over.
Both clips were generated on this page's free engine — the second continues the first from its exact final frame, using the same upload-and-prompt flow this page hands you. Every output is AI-generated, so your seam and your scene will vary. Watch the handoff, not the boat.
A paper boat drifts down a rain-water stream in golden light, and after six seconds the clip simply ends — mid-drift, the way every generated clip ends. This is the frame the continuation anchors on: the calmer that final frame, the cleaner the pickup.
Generate a first clipThe final frame went back in with a continuation prompt: keep drifting, pass the drain grate, spin once in the eddy. The stream, the light and the boat carry over because the continuation starts from the exact pixels the first clip ended on — that is the whole trick, made visible.
Extend your own clipIt is the answer to the most common complaint about generated video: the clip is good and then it just stops. A video extender takes the last frame of an existing clip as its starting image and generates the seconds that follow, so the scene, the light and the subject carry over instead of resetting. Chain the trick and a six-second clip becomes twelve, eighteen, as long as the story needs.
The technique is honest and worth understanding, because it is what every such tool actually does under the hood: video models generate forward from an image, so "continuing" means handing the model your clip's final frame and a prompt for what happens next. On this page that flow is explicit — export the last frame of your clip (a paused screenshot works), upload it in image-to-video mode, and describe the continuation.
Explicit beats hidden: you choose the exact frame to continue from, and you can steer the next seconds instead of hoping.
The pair on this page is our own run: a paper boat drifting down a rain stream, generated once, then continued from its final frame with a prompt asking the boat to keep drifting and spin once in an eddy. Watch the cut point — the scene, the light and the boat carry across it.
The honest limit shows there too: the tool preserves what the last frame can see, so motion that is mid-blur at the cut can stutter, and off-screen elements are re-imagined rather than remembered. Pick a calm frame to continue from and the seam disappears; pick a chaotic one and it shows.
Creative engine
Last frame in, continuation prompt written, next seconds out — repeat until the story is told.
The continuation starts from your clip's exact final frame — same scene, same light, same subject, no reset.
The new seconds follow a new prompt, so you can hold the motion or turn the scene while keeping the world.
Continuation pairs naturally with the other camera tricks here: freeze the peak with the bullet time effect and carry on past the release, keep a dance running longer with the twerk generator, or start from a still photo in the kiss video generator and stretch the moment it ends.
From a clip that ends too soon to one that ends on purpose.
Pause your clip on its final frame and export or screenshot it. Choose a frame where the subject is clearly visible and motion is settled — the model continues what the frame shows, so a clean anchor beats a blurry one.
Switch to image-to-video, upload the frame, and write the continuation the way you would direct the next shot: what keeps moving, what changes, where the camera goes. Repeat the scene's look in words — same light, same setting — to hold continuity.
Run the extension and cut it onto the original in any editor. Need more? Take the new clip's last frame and go again — each pass adds seconds while the scene carries forward. If an extension pass fails, its credits come back on their own (content-policy blocks are the one exception).
Anyone whose best generation ended four seconds before the idea did.
A six-second cap forces scenes to end mid-thought. Chaining extensions lets a generated scene run to its natural beat — arrival, action, aftermath — instead of the model's timer.
When one generation finally nails the look, re-rolling is a gamble. Continuing from its last frame keeps the take you fought for and buys the extra seconds the edit needs.
Common questions about continuing clips on Kavel.
Just the final frame. The generator takes an image as its starting point, so the flow is: pause on the last frame, screenshot or export it, upload, and prompt the continuation. It is one honest step more than a magic button — and in exchange you choose exactly which frame the continuation grows from, which is where most of the control lives.
The starting pixels match perfectly — the continuation begins on your exact final frame. From there the model re-imagines motion, so the match stays strong for scene, light and subject, and weakest for fast motion crossing the cut and for anything off-screen, which the model never saw and must invent. A settled anchor frame and a prompt that restates the scene's look give the cleanest seam.
There is no hard cap — each extension's last frame can seed the next. Drift is the real limit: tiny changes compound with each generation, so scenes hold together well for two or three chained passes and get progressively looser after that. For long pieces, plan cuts where a scene change is natural and start fresh anchors there.
Yes — that is half the point. The continuation follows your new prompt, so from the same anchor frame you can keep the motion going, introduce something new, or swing the mood. The world carries over; the direction is yours.
Each extension is a normal generation on Kavel's free video engine, and new visitors start with free credits. The estimate shows before each run and a failed generation is refunded automatically (unless it broke the content policy) — so a twelve-second story costs two short runs, not an upgrade.
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. |
The tricks on either side of an extension, and the studio they run in.
Upload the last frame, describe what happens next, and the scene continues instead of starting over.
