Overview
minia.art is an AI pixel art generation platform built by Ketryon, a software studio based in Eskilstuna, Sweden. It is not a general-purpose image generator and makes no real attempt to compete with the Midjourneys of the world on breadth. Instead, it commits to a single aesthetic — pixel art — and then stretches that aesthetic across an unusually wide range of output formats: hero banners, social cards, sprite sheets, tileable textures, icon sets, transparent PNGs, and short looping MP4 clips.
The problem it addresses will be familiar to anyone who has shipped a side project, run a social account, or managed an ad set. Visual assets are a bottleneck. Commissioning an illustrator is slow and expensive; stock libraries rarely carry pixel art that matches a specific brief; and general AI image models tend to drift toward painterly or photographic output rather than the crisp, deliberately low-resolution look the pixel aesthetic demands. minia.art positions itself between those options — faster than commissioning art, more stylistically disciplined than prompting a general model.
The audience is narrow and clearly stated: product teams needing hero art and in-app scenes, creators needing avatars, banners and thumbnails, and marketers producing ad creative in every aspect ratio. Pricing targets individuals and small teams rather than enterprise design departments, with a free tier of twenty generations per month and no credit card requirement. The category is crowded with generic generators; the specificity of the pitch is the differentiator. The thesis of this review is that this specificity is both the product's greatest strength and the hard boundary of its usefulness.
Key Features
- Eighteen distinct art styles. Rather than funnelling every request through a single model, minia.art exposes a style carousel covering scenes, isometric views, retro looks, cartoon styles, sprite sheets, item sheets, character models, icon sets, and block textures, among others. Preview thumbnails let users see the aesthetic before spending a generation. This matters because pixel art is not one look — a 16-bit RPG sprite and an isometric diorama share almost nothing stylistically, and collapsing them into one model would degrade both.
- A node-based canvas with connection awareness. This is the most technically interesting element of the product. Users drop images, notes and video clips onto a board and wire them together, and those connections carry semantic weight: a note becomes creative direction, an image becomes a reference, and a clip stands in for the still frame it originated from. The interface surfaces what each node has picked up before generation runs, which removes much of the guesswork typical of multi-reference prompting.
- Text-to-pixel-art generation. The baseline workflow is one sentence in, one finished asset out. minia.art advertises results in seconds with no palette selection, canvas configuration or art training required. Sample prompts on the landing page — a tiny astronaut floating past a ringed planet, a view through a rainy window over a city at night — signal the intended level of description: evocative but not exhaustive.
- Photo-to-pixel-art conversion with adjustable strength. Users upload an existing photograph and dial a strength slider governing how much of the original likeness survives the conversion. A separate palette image can be supplied to steer the colour scheme. The strength control is a practical addition, since a fixed conversion would produce either unrecognisable output or something indistinguishable from a simple filter.
- Motion and looping video generation. Available on Pro and Studio tiers, this runs a video model to turn a still into a five- or ten-second looping MP4, with a choice of ambient, camera-drift, character or action motion. A prompt field describes what should move. The output suits reels, stories and hero sections where short, subtle motion outperforms a static image.
- Sprite and character animation. Walk cycles, idles and visual effects can be generated from a text prompt or from art already created on the platform. Results download as a looping GIF for sharing or as a sprite sheet for direct import into a game engine, which removes a conversion step that usually requires a separate tool.
- Seamless tiling, transparency and upscaling. A single toggle makes image edges wrap so textures repeat without visible seams, useful for backgrounds, patterns and surfaces. A separate toggle produces transparent PNGs, and assets can be upscaled up to 8× without softening the pixel edges — an important detail, since naive upscaling destroys the crispness that defines the style.
- Batch generation with seed locking. Users generate four or ten variations of a concept in a single run and choose from a carousel rather than re-rolling one at a time. Locking a seed holds the composition steady while the prompt changes, which is how coherent sets of related images get built.
How It Works
Onboarding is deliberately short. A visitor lands on the homepage, sees a live prompt example rendered as pixel art, and is offered two calls to action: start creating for free, or browse the examples gallery. Signup requires no credit card, and the free tier grants twenty generations per month on an indefinite basis.
Once inside, the workflow follows three documented steps. First, the user describes an idea — typing a prompt or uploading a photo or palette to steer the result. Second, they tune the details: selecting one of the eighteen styles, choosing an aspect ratio, setting batch size, toggling transparency or tiling, configuring the upscale factor, and optionally locking a seed. Third, they download the finished asset as a crisp PNG, GIF or MP4. Everything remains in a generation history for later retrieval.
The learning curve is shallow by design. The landing page explicitly promises no onboarding maze, and the interface reflects that: the tool is organised around a generate panel, an animate panel and a canvas view rather than a complex project hierarchy. The main conceptual hurdle is the node canvas, which requires users to understand that connections between nodes carry meaning. That is an advanced workflow for building scenes iteratively, and it is entirely optional for anyone who simply wants to type a sentence and download a sprite.
A dashboard mockup on the homepage indicates the product tracks generation history, credit usage and asset collections, with a percentage indicator suggesting a monthly quota meter. Animated loop generation is described as taking a minute or two, and users are told they can leave the page and return — an honest disclosure about video-model latency that most tools of this type omit.
Use Cases
A solo indie developer prototyping a 2D game. Before a project is presentable, it needs walk cycles, idle animations and item icons. Rather than commissioning a pixel artist for placeholder assets, the developer generates sprite sheets through minia.art, exports GIFs for the devlog, and swaps in final art later. Seed locking keeps the same character consistent across every animation in the set.
A SaaS founder preparing a launch. A founder with a retro-leaning brand needs hero art, an OG card for social sharing, and blog cover images. They generate a wide-ratio scene, batch ten variations, pick the strongest, and export a transparent PNG to overlay on the marketing site — all without opening a design tool or paying for a stock subscription.
A content creator maintaining a visual identity. A streamer or YouTube creator wants a profile picture, channel banner and thumbnail art that reads as intentional rather than templated. Photo-to-pixel conversion turns an existing portrait into a recognisable pixel avatar while preserving facial features, and the tiling toggle produces repeating backgrounds for stream overlays.
A performance marketer running creative tests. An agency marketer needs multiple ad concepts in square, vertical and horizontal ratios. Batching generations and exporting at scale produces enough variations to test properly instead of running a single asset on faith. The one-minute animation capability adds short video creative, though the ten-per-month cap on Pro limits how far that testing can extend.
A hobbyist building a personal site or community. Someone who wants a distinctive avatar, a pixel logo or a decorative background can use the free tier at no cost. Twenty monthly generations cover a handful of assets indefinitely, and free utilities such as the image pixelation tool extend the platform's value beyond generation itself.
Design & User Experience
The interface reads as modern and restrained, with the pixel aesthetic carried through the marketing site rather than quarantined to the product. Navigation is shallow — generate, animate, canvas, history — and the homepage does an effective job of demonstrating output before asking for a signup, which is the correct order of persuasion for a visual tool.
The standout design decision is the node canvas. Connection-aware nodes are a genuinely differentiated idea, and showing users what each node contributed before generation runs addresses the biggest frustration in multi-reference prompting. It is also the sharpest edge in the product: node graphs impose a mental model that casual users will not have, and the cognitive jump from single-prompt generation to wired canvases is steep. The optionality mitigates this, but the two modes can feel like different products sharing a shell.
Elsewhere the experience is clean if conventional: style carousel, aspect-ratio selector, batch controls, download buttons. Mobile responsiveness is not fully evidenced from the marketing site, and a multi-node canvas is inherently a desktop-first interaction. The honest latency disclosure around video generation is a small but appreciated touch.
Pricing & Value
Pricing is publicly documented on the pricing page and splits into three tiers.
- Free — twenty generations per month, free indefinitely, no credit card required. Best for hobbyists and anyone evaluating the tool.
- Pro — roughly $10 per month, billed in SEK, including 1,000 generations per month, 10 looping videos, all eighteen styles, sprite animations and commercial use rights. Best for indie developers and solo creators shipping regularly.
- Studio — roughly $36 per month, including 5,000 generations and 30 looping videos, plus everything in Pro. Best for small teams and agencies producing at volume.
Annual billing saves around 17 percent. The free tier is genuinely usable rather than a demo, and the paid tiers undercut general-purpose image subscriptions that neither specialise in pixel art nor bundle looping video at these volumes. The main friction is currency: SEK-denominated pricing means international buyers may see small variations depending on exchange rates, which complicates budgeting.
Who Is minia.art Best For?
Indie game developers and small studios. Anyone building a 2D title who needs sprite sheets, walk cycles and item icons without hiring a pixel artist for every asset will find the sprite-animation and seed-locking workflow directly aligned with how they actually work. The export paths — GIF for devlogs, sheets for engines — reflect an understanding of that pipeline.
Retro-styled creators and streamers. People whose brand leans on pixel aesthetics need avatars, banners, thumbnails and overlay textures on a recurring basis. Doing that consistently through one tool is cheaper and faster than commissioning individual pieces.
Marketers producing multivariate creative. Teams testing multiple ad concepts across ratios benefit from batch generation and transparent PNG export, provided their video needs stay modest.
Two groups should look elsewhere. Anyone needing illustration outside the pixel aesthetic will find nothing here — a general model or a human illustrator is the right call. And teams planning heavy video creative testing will hit the ten-video monthly cap on Pro quickly; a dedicated video generation tool would serve them better.
