No game session · No screen recorder · Nothing to install
Get the blocky footage without playing the run. Describe the course and the camera, and the clip is generated for you.
A real clip generated on Kavel with the same model this tool uses, from a single description of the run. Floating voxel islands over open sky, blocky trees, a stone path underfoot, hands in the frame and the hotbar along the bottom — none of it recorded, none of it built. Every output is AI-generated, so your own clip varies with the prompt you write; read this as a guide to what the tool does, not a fixed result you will get back.
One description of the course went in and this came back — voxel islands hanging over open sky, blocky trees, a stone path underfoot, hands in frame and the hotbar along the bottom. Nothing was built and nothing was recorded. It is also where the limits show: pause it and the blocks do not sit on a perfect grid and the hotbar is decorative rather than functional. In motion, which is how background footage is watched, neither reads.
Open the video studioThe same tool, a different shot: the camera tracks behind and slightly above a character vaulting up a tower of coloured blocks, with the voxel landscape falling away below in late-afternoon light. Changing the camera line in the prompt is what produced this — the scenery description barely moved. That is the lever most people leave untouched, and it is the difference between footage that feels like gameplay and footage that drifts.
Open the video studioIt writes gameplay-style footage from a description instead of capturing it from a session. You say what the run looks like — floating blocky islands over a void, a first-person view with the hands in frame, a fast forward camera, bright daylight — and get a short clip with that content, in the flat-textured cubic style people recognise. Nobody has to build the course and nobody has to land the jumps.
The reason this exists as a search is the background-footage economy: satisfying block-world clips are one of the most-used backdrops for text-over-video posts, and getting them normally means either downloading someone else's recording, which everybody else is also using, or building a map and recording your own runs, which is hours of work for thirty usable seconds.
Generation sits between the two: original footage, minutes of effort. On Kavel it runs on Veo 3.1 in the browser, with free credits to start. Describe the run, pick the format, and the credit estimate updates before you commit; a failed run is refunded automatically (unless it broke the content policy). Be honest with yourself about what this is, though.
It produces something that reads as gameplay at a glance and in motion — which is exactly what background footage needs — but it is not a recording of a real session, the physics are described rather than simulated, and a viewer who pauses on a frame will see it. For a backdrop behind text, that does not matter. For anything presented as an actual play session, it does.
Creative engine
Describe the run and the camera, pick the format, and the credit estimate updates before you generate.
Writes gameplay-style parkour footage from a description instead of capturing a real session.
Puts the camera under your control — first person, chase view, speed and lighting are all prompt.
Produces original clips rather than the same downloaded background everyone else is using. Parkour footage comes from a description, not an upload. To put your own photo into that world, the Minecraft photo generator rebuilds it from cubes and the pixel art converter flattens it to a sprite.
From a sentence to usable footage, with the cost shown before you commit.
Say what the player is crossing — floating islands over a void, a tower climb, a path through blocky forest — and name the light. Specific beats grand: one clear course reads better than an epic nobody can follow.
This is the part people skip and then wonder why the clip feels wrong. Say first person with hands in view, or a chase camera behind the player, and say how fast it moves.
Generate, then grab the clip for the edit you had in mind. A failed run is refunded (unless it broke the content policy), so a second take of the same course costs nothing but the wait.
Faceless channels, background filler, a creator with no capture setup.
The biggest use by far. Satisfying motion under a caption holds attention, and original footage does not carry the staleness of a clip a thousand other accounts already used.
A few seconds of movement to cut between talking segments, without interrupting your edit to go and record something.
Pull a frame for a thumbnail, or visualise a course idea before deciding whether it is worth building for real.
Common questions about generating blocky gameplay footage with Kavel.
No, and you should not present it as such. It is generated footage that reads as gameplay in motion, which is what makes it work as a backdrop, but no session was played and no map exists. The physics are described rather than simulated, so a frame-by-frame look will not hold up the way a screen recording does. Use it as visual texture behind your content, not as evidence that a run happened.
Short — a few seconds, the length these models produce in one pass. That is the natural length for a looping backdrop anyway. If you need a longer stretch, generate two or three takes of the same course and cut them together; because they are generated rather than recorded, the joins are easier to hide than they would be between two real runs.
Yes, and it is the highest-leverage thing you can put in the prompt. Say first person with the hands visible, or a chase camera behind the player, and describe the speed and whether it is smooth or handheld. Prompts that describe only the scenery tend to come back with a camera that wanders, which is what makes a clip feel off.
Because it was written rather than rendered by a game engine. Blocks may not align perfectly on a grid, the hotbar is decorative rather than functional, and details shift between frames. In motion — which is how the footage is actually watched — this is close to invisible. Paused and studied, it is obvious. Judge it at playback speed, because that is how it will be seen.
Read the terms of service for what you can do with generated output, and use judgement about the platform: many will treat unlabelled AI footage differently from a recording, and some require disclosure. This tool is also not affiliated with or endorsed by Mojang or Microsoft — it produces artwork in a blocky voxel style, and Minecraft is a trademark of its respective owner.
Not faithfully. The generator works from a description, so it produces a plausible course matching your words, not a reproduction of a build you already have. If a specific map matters — showing your own creation, for instance — record it. Generation is for when any convincing course will do, which is most background-footage jobs.
The estimate appears before you generate and moves with the model, resolution and format you pick, so the cost is visible first. Video costs more than an image, which is worth knowing before you queue several takes. Failed runs are refunded automatically (unless it broke the content policy), and new accounts start with free credits.
New accounts start with free credits, which is enough for a first clip without paying anything. Beyond that, generating video has a real compute cost on every provider, so anything claiming unlimited free AI video is either capping you somewhere else or watermarking the output. The cost is shown here before you press generate rather than after.
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. |
Other ways to get footage without filming it, and the blocky look on stills.
Describe the course and the camera, check the credits, and get a blocky parkour clip in a single pass.
