Product campaign
Product, headline and CTA relocate; decoration degrades first.




Responsive Visual Asset
Author one structured creative. RVA recomposes it for the space available while preserving layout intent, hierarchy and protected content.
Created by Airovo Technologies.
Drag the right edge, or pick a ratio — the asset genuinely re-resolves.
The problem
Modern visual content has to work across an expanding number of shapes. Today that means manually adapting the same idea again and again — because existing image formats preserve pixels, not composition intent.




Four fixed exports of the same idea — each cropped and checked by hand, maintained separately.

master.rva



Author once, declare intent, then validate and export conventional assets for the placements you actually need.
Why we built RVA
We saw a gap between how visual content is made and how it has to live: an expanding number of shapes, surfaces and contexts — while the formats we ship preserve only pixels. So we researched, designed a format, and engineered it end to end.
RVA is the result of engineering inquiry, not a marketing invention: a structured visual asset, a deterministic resolver, SDKs across runtimes, and a Studio to author it.
How RVA thinks
A small, legible pipeline turns one creative into any shape without losing design intent.
Resources become semantic elements — images, vectors, text, groups.
Roles, focal & protected regions, constraints and priorities declare what matters.
rva-core selects a topology and resolves exact geometry for the viewport W×H.
The same resolved scene reaches web, native or server renderers.
Design intent, preserved
RVA keeps the parts of a composition that matter. Regions, topologies and priorities are declared once and honoured everywhere.
Tell RVA which part of an image must survive cropping and scaling. Protected regions are hard constraints.
Composition rules switch by aspect ratio: a wide banner and a tall story are authored as separate topologies.
When space is tight, low-priority decoration hides first — the hierarchy and the subject stay intact.
Built for creative production
Declare the placements your campaign must serve. RVA validates them and exports conventional assets — so it fits existing production workflows without waiting for platforms to natively accept .rva.
ONE RESPONSIVE MASTER
Presets reflect configured production targets and may change with platform requirements.
Author
one responsive master in RVA Studio
Preview
across declared targets
Validate
solve every placement
Export
a conventional PNG bundle
RVA Studio
Studio is the professional authoring environment: build the semantic scene, declare responsive intent, then validate and export.

Scene & Assets
Semantic elements and imported resources, with roles and grouping.
Topology authoring
Author a composition per aspect-ratio range, side by side.
Constraints & regions
Gaps, alignment, containment, focal and protected regions.
Preview, validate, ship
Declared targets, cross-size validation, Save .rvap / Export .rva.
macOS · Apple silicon · 10.5 MB
.rvap — editable project.rva — distribution asset
RVA Viewer
A lightweight, read-only viewer for Responsive Visual Assets. Drop in a .rva, resize the window, and see the same composition the resolver produces everywhere.

macOS · Apple silicon · 7.9 MB
Engineering proof
RVA is implemented infrastructure: a portable container, a deterministic resolver shared by every runtime, and adapters that prove it across web, native and server.
Adapter surface
The RVA ecosystem
RVA is a set of interoperating pieces — a core, tools, SDKs and a specification — held together by cross-runtime conformance.
A deterministic resolution engine: parse, validate, select a topology, resolve geometry.
The professional authoring environment for scenes, topologies, targets and export.
A lightweight viewer to inspect any .rva and watch it recompose as you resize.
Web, native and server integration over one shared core and one primitive surface.
A portable container and model: structured scenes, topologies and semantics.
Cross-runtime fixtures that hold Rust, WASM and the C ABI to identical output.
CTA Regions
A CTA Region is semantic geometry: RVA resolves where it is for every viewport and exposes a stable id. Your application maps that id to an action — no scripts or behavior live in the asset.
Two labelled regions. Behavior is supplied by the host, not embedded in the file.
// Web family
el.ctaHandlers = {
"shop-now": () => go("/product"),
"learn-more": () => go("/details"),
};
// Any host: resolve, then bind the ids it wants
scene.ctaRegions
.filter((region) => region.visible)
.forEach((region) => bind(region.id, region.normalizedBounds));Open technology
RVA is open technology: a published specification, an MIT-licensed core and runtime ecosystem, conformance fixtures and examples — all in one public repository.
The portable container and structured scene model, versioned with the code.
Read the specThe deterministic resolver and its runtime families (Rust · WASM · C ABI).
Browse the coreRVA was created by Airovo Technologies.
Examples
The same asset in each category, recomposed across aspect ratios by the same engine. Notice the subject, headline and call-to-action hold their intent as the frame changes.
Product, headline and CTA relocate; decoration degrades first.




A protected focal subject survives every crop.




Composes across wide banners, squares and tall stories.




RVA was built by Airovo
Airovo is a technology engineering company. We create original software technologies and collaborate with companies to turn ambitious ideas and emerging technologies into real products.
Original technologies and products, from first principles.
Complex software systems, architecture through production.
Alongside technology companies, agencies and enterprises.
Build it with the engineering team behind RVA.