Native UGC galleries inside your iOS or Android app: the SDK, not a standalone Idukki app
Idukki does not ship an installable app of its own. What we ship is a toolset your own iOS or Android app calls to render native UGC galleries: a shipped, tested React Native package, a tested Android client, and an iOS client that is honestly still pre-release.
Your CTO asks on the vendor call: is this a native SDK or a dressed-up WebView? For a long time the honest answer was the second one. For React Native that has changed. Android and iOS sit at different stages, and this piece says exactly where.
In this article
Most "do you have a native SDK" questions are two questions folded into one. Does content render as native UI instead of a web view? And is Idukki asking a merchant to build or promote a second app? The first answer is now yes for React Native, with Android and iOS following. The second has always been no. Staying no is a deliberate product decision.
Two ways to be "in the app," and they are not the same claim
Before this SDK existed, "does the widget work in native apps" had a real but limited answer, shipped as part of the widget speed program: a tap bridge that fixes the two things that silently break a web widget inside a WebView (product taps going nowhere because WebViews drop window.open, and storage degrading ungracefully when a WebView ships with it disabled). That bridge is real, shipped, and stays the right answer for any app that already renders its own site, or a widget page, inside a WebView. It is not what "native rendering" is asking for, and we try not to describe it as one.
Which path fits your app today
Start here
Does your app already render your site, or a widget page, in a WebView?
- Yes
WebView tap-bridge
Shipped. It fixes product taps and storage inside the WebView. Typical integration takes days.
- No, the app is native
Native client SDK
Pick by stack. Each one is at a different stage.
- React Native: Shipped and tested. Ready for a design partner.
- Native Android: The tested data and pagination controller works now. The Compose UI exists as source and has never been compiled.
- Native iOS: Swift source only, never compiled. Tell us if this is your blocker.
WebView tap-bridge
Your app already renders a page (yours, or a widget URL) inside a WebView.
Wins at
- No native build required on your side
- Ships in days, not weeks
- Diagnostic panel included (idk_appcheck=1)
Struggles with
- Still a WebView underneath: no native scroll physics, no native video player
- Reads as an obviously embedded web strip to a design-literate user
Native client SDK
Your app is React Native today, or fully native Swift/Kotlin and a WebView reads as visibly bolted-on.
Wins at
- Real native UI: FlatList/LazyVGrid/Compose, not an iframe
- Native scroll, native gestures, native video playback
- Same product-tagging + add-to-cart contract as the web widget
Struggles with
- A real integration, not a script tag: you write code against the SDK
- Android ships data-layer only today; iOS is pre-release
Why a toolset, not an app
A UGC gallery is a feature inside someone else's app. Nobody opens an app to browse Idukki. They open a merchant's app to shop. The gallery earns its place inside that experience. It sits next to the merchant's own cart, accounts, push notifications and App Store listing. "The Idukki app" would ask shoppers to install a second app for one feature. It would compete for a home-screen slot the merchant already owns. That is worse for the merchant. It also breaks with how the web widget works. That runs on the merchant's own domain, under their brand. We host no page of our own for it.
So the native mobile work follows the same shape as the web SDK, not a new shape: a client library and a UI component (or, for Android today, just the client) that a merchant's engineers drop into an app they already own and control. The cart stays theirs (an onAddToCart delegate, never our checkout). The analytics stay theirs (an onEvent callback mirrored into whatever mobile analytics SDK they already run). The App Store listing stays theirs. We ship the gallery; they ship the app.
What's shipped: React Native
The React Native package renders the gallery as a real <FlatList>-backed grid. There is no WebView or iframe in the tree. Tapping a product opens a native modal with tagging. No browser sheet appears. It reads the same public gallery endpoint as the web widget. So there is no second content pipeline. Tag a product once and it shows everywhere.
import { IdukkiGallery } from '@idukki/widget-sdk-native';
<IdukkiGallery
layoutGuid="your-gallery-guid"
businessGuid="your-business-guid"
onAddToCart={async ({ productId, variantId, quantity }) => {
// Call your own app's cart API here. Return { ok: true } on
// success — same contract the web widget's cart hook uses.
await myAppCart.add(productId, variantId, quantity);
return { ok: true };
}}
onEvent={(event) => {
// Mirror into Firebase / Amplitude / whatever you already run.
analytics.logEvent(event.event, event);
}}
/>That is the whole integration for a grid layout: a component, a cart callback, an optional analytics callback. Behind it, the client resolves a persistent per-device identity (there is no browser cookie jar on native), fetches and paginates the gallery feed, and maps the six theme tokens the dashboard's Brand Kit already produces onto the component's styling, so a merchant's colours and radius carry over without them touching a design file.
Two things it does not do yet, said plainly now. Only the grid layout ships: no reel, carousel or story layout in v1. And the gallery component is reviewed and unit-tested. No production RN app has mounted it yet. It is ready for a design partner. A general-availability npm install comes later.
What's shipped: Android
The Android client is a plain Kotlin/JVM port of the same logic. URL construction, pagination fallback and theme-token reconciliation are identical. So the web, React Native and Android clients cannot drift on what the backend sends. On top sits a pagination and loading-state controller. A UI layer drives a gallery screen with it: load more, refresh, loading and errors. It is also plain Kotlin, also fully tested.
val client = IdukkiNativeClient(
IdukkiNativeClientConfig(
layoutGuid = "your-gallery-guid",
businessGuid = "your-business-guid",
fetcher = HttpGalleryFetcher(),
)
)
val controller = GalleryController(client, scope = viewModelScope)
controller.load()
// controller.state: StateFlow<GalleryUiState> — posts, theme,
// hasMore, isLoading, error. Render it with Compose or any UI layer.A Jetpack Compose UI layer on that controller now exists too. It has a grid, a tap-to-open product sheet and add-to-cart wiring. It was cross-checked field-for-field against the React Native component. It has never been compiled, said as plainly as the iOS caveat below. This sandbox's network policy blocks the Android SDK download, as it blocks Swift's. We tested that directly. The data and pagination layer underneath is fully tested. Only the rendering code is unverified. If the tested controller is enough for your Android app, it is usable today.
iOS: written, not yet a claim we make
The honest state of iOS: Swift source exists and has never been compiled. It mirrors the Android and React Native clients field-for-field, against documented Foundation and URLSession APIs. That is the actual current state. We would rather say it plainly here than have a sales conversation discover it. An Xcode build is the next step, once a design partner's stack needs it.
How this gets tested, and how you could test it yourself
Every claim above about "tested" means an automated suite that runs against a mocked network, by design: neither client's core logic should depend on Idukki's production infrastructure being reachable just to prove the URL-building and pagination logic is correct. Both the React Native and Android clients also carry a second, opt-in test that hits the real public gallery endpoint for a true end-to-end check, kept separate and disabled by default so a normal test run never needs network access.
- React Native:
npm testruns the full mocked suite;IDUKKI_LIVE_TEST=1 npm run test:liveruns the real one. - Android:
gradle testruns the full mocked suite;IDUKKI_LIVE_TEST=1 gradle test --tests "*LiveIntegrationTest"runs the real one. - iOS:
swift buildandswift testare the two commands that would turn the "never compiled" caveat above into an actual result, the first time anyone with Xcode access runs them.
None of the three packages is public yet: no npm, Maven Central or Swift Package Index. They ship today as source to design partners. Other unfinished parts of the platform ship the same way. Is your app React Native, Android or native iOS, with a WebView tile that looks wrong? Talk to us directly. Do not wait for a public release.
FAQs
Does Idukki have a native iOS and Android app?
No, and that is intentional. Idukki does not ship a standalone, installable app for shoppers to download. What we provide is a toolset (SDKs and, where relevant, a WebView bridge) that a merchant's own iOS or Android app uses to render UGC galleries as a feature inside that app, the same relationship the web widget already has to a merchant's own website.
What is the difference between the WebView bridge and the native SDK?
The WebView bridge fixes the two things that break the existing web widget inside a WebView (dead product taps, unreliable storage) so it works correctly there. It is still a web view underneath. The native SDK renders the gallery as real platform UI (a FlatList in React Native, a Compose grid on Android once that ships) with no WebView involved at all.
Is the React Native SDK on npm?
Not yet. It is shipped as source to design partners today, consumed as a local dependency, while the API stabilises. It has a full automated test suite (30 tests) and is ready to integrate; it is not yet a public, self-serve npm install.
Is there a native iOS SDK?
Swift source exists and mirrors the Android and React Native logic. It has never been compiled or run. There is no verified iOS SDK yet, only written source waiting on a first real build. Tell us if this blocks your integration. It changes how we prioritise it.
Can I use just the Android client without a UI component?
Yes. The data and pagination controller works on its own, with 25 automated tests. It covers gallery fetch, pagination, theme-token mapping and loading/error state. You can render the grid with your own components using the state it exposes. A Compose UI layer on top now exists as source. It has never been compiled, so we would not call it finished.
Sources & notes
- 1Inside the widget speed program: cutting our embed JS by two thirds · Covers the shipped WebView tap-bridge this article builds on.
- 2API platform · Where this SDK sits relative to the rest of the platform roadmap, staged honestly (live / private beta / roadmap).
- 3Note · Test counts (30/30 React Native, 25/25 Android) are current as of this article's last update. So is the "iOS never compiled" / "Android Compose UI never compiled" status. Both will move as the SDKs mature. Treat them as a snapshot.
More from Rohin Aggarwal
- Embedding
How to make your Instagram feed shoppable on your own site
A shoppable Instagram feed puts your Instagram posts on your site, tagged with your products. Seven steps from connecting the account to counting orders, and how it differs from Instagram Shopping.
- Embedding
How to Embed Instagram Stories on Your Blog (The Honest Way)
Stories expire in 24 hours and Instagram's embed tools don't cover them. What actually works: capture with permission, display as a widget, and the API rules behind it.
- Embedding
How to Embed an Instagram Feed on Your Website (2026 Guide)
Every working way to embed an Instagram feed on a website in 2026: single-post embeds, the Instagram Graph API after the Basic Display shutdown, hashtag feeds, and shoppable UGC galleries.