One photo in · Retro sprite out · Nothing to install
Convert a picture to pixel art in one pass. Your own photo, redrawn as a 16-bit sprite instead of blurred into big squares.
A real sprite conversion from this page’s model. Drag the divider: the same tabby cat, the same upright pose, the same pale-blue background — now drawn on a coarse grid with flat colour blocks and hard aliased edges. Look at the whiskers and the eyes: they survive as deliberate single-pixel lines, which is exactly what a pixelate filter would have destroyed. Every output is AI-generated, so your own result varies with the photo you upload — read this as a guide to what the tool does, not a fixed result you will get back.
Drag to compare before and after
It rebuilds your photograph in the visual language of a retro video game: a coarse grid of large square pixels, a deliberately small palette of flat colours, hard aliased edges and stair-stepped diagonals, with no gradients or soft blur anywhere. The subject stays recognisable — your cat is still your cat — but every surface is now made of discrete blocks of colour.
This is not the same as pixelating an image, and the difference is the whole reason this tool exists. Pixelating averages the pixels you already have into bigger squares, which destroys information: faces turn to mush, thin features vanish, and the result reads as a censored photo rather than as artwork.
Converting redraws the subject for the low-resolution grid, deciding which shapes survive and which get simplified, the way a sprite artist would. That is why the eyes, whiskers and stripes in the example below are still legible at a size where a blur filter would have erased them. It is also not the same as a Minecraft block build: this output is flat 2D art, not cubes with volume.
The conversion runs on the Nano Banana image model, in the browser, with free credits to start. Upload the source art — the flow starts in image-to-image — pick the format, and the credit estimate updates before you commit; a failed run is refunded automatically (unless it broke the content policy). Two things decide how good the result looks.
Start from a photo with a clear subject and good contrast against its background, because the grid has very few pixels to spend on separating one thing from another. And say what the sprite should look like: name the era or resolution you want, since a chunky 16-bit look and a fine-grained modern one are different requests and the model will not guess which you meant.
Creative engine
Upload a photo, describe the sprite look you want, pick the format, and the credit estimate updates before you generate.
Redraws the subject for a coarse grid instead of averaging your pixels into larger squares.
Produces flat 2D sprite art with hard edges and a limited palette — no gradients, no blur.
Works in the browser with free credits — no sprite editor, no palette setup, no grid maths. Pixel art is flat by definition — the Minecraft photo generator rebuilds the same picture from cubes with real depth, and the parkour video generator puts that blocky look in motion.
From one photograph to a finished sprite, with the cost shown before you commit.
One clear subject that separates cleanly from its background. The coarse grid has very few pixels to spend, so contrast does the work detail cannot.
Say which era or resolution — a chunky 8-bit sprite, a 16-bit game character, a fine modern pixel illustration — and mention the palette if the colours matter.
Generate, then export the sprite at the size your project needs. A failed run is refunded (unless it broke the content policy), so trying a coarser or finer version costs nothing but the wait.
Game jams, itch.io pages, avatars that should look 16-bit.
A sprite version of your own photo reads well at small sizes and stays distinctive where a cropped photo turns into a blur.
Turn reference photos into placeholder sprites fast when you are prototyping and do not have an artist yet.
A flat, limited-palette grid is the starting point for perler beads, cross-stitch and pixel posters — the art step, before any charting tool.
Common questions about turning a photograph into flat retro sprite art with Kavel.
A pixelate filter averages the pixels you already have into bigger squares, so information is destroyed: faces smear, thin features disappear, and the result looks censored rather than drawn. This redraws the subject for the coarse grid, choosing which shapes to keep and which to simplify. That is why fine features like whiskers or the line of an eye can survive here as deliberate single-pixel marks, and cannot survive a blur.
You describe both in words rather than setting them numerically. Ask for a chunky 8-bit look, a 16-bit sprite, or a fine modern pixel illustration, and name a palette if the colours matter — muted earth tones, a Game Boy green, four colours only. Treat it as art direction rather than an exact pixel count; if you need a literal 32×32 grid, do the conversion here and resize afterwards.
No. This is flat 2D sprite art on a grid. A Minecraft-style rebuild makes the scene out of cubes that have volume and cast shadows, which is a 3D look — we have a separate tool for that. Pick this one for avatars, sprites and craft patterns; pick the block one for anything meant to look like the game.
Yes, and a portrait is one of the better subjects because the composition is simple. Expect the likeness to be suggestive rather than exact: a coarse grid has too few pixels for a face to stay literally identical, so what carries over is hair colour, glasses, the shape of the head and the expression. That is the medium doing its job, not a failure of the conversion.
Busy scenes with several competing subjects, low-contrast pictures where the subject and background are similar in tone, and images that rely on fine texture — fur detail, lace, foliage. All three ask the grid to represent more than it can hold. Crop tighter, or pick a frame with a cleaner silhouette.
The terms of service set out what you can do with images you generate — read those before shipping the result in a product or a store listing. Also remember that the input matters: converting someone else's photograph or a copyrighted character does not make the result yours to sell.
The estimate shifts with the format and resolution you select, so the cost is visible first. Failed runs are refunded automatically (unless it broke the content policy), and new accounts start with free credits — enough to try a couple of looks before spending anything.
Not directly. This produces the artwork, not a chart with a colour key and a stitch count. It is the step before that: convert the photo here, then feed the flat image into whichever charting tool your craft uses, which is far easier than charting straight from a photograph.
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 restyle a photo you already have, and the model behind the edit.
A real sprite, not a photo blurred into big squares. Name the palette if the first pass reads too modern.
A real edit made with this tool — drag your own photo in on the left, the prompt is already written, then press Generate.
