Capsora is a whole AI wardrobe app template constructed with Flutter, GetX and Materials 3. {Photograph} a garment and a imaginative and prescient mannequin names it, colours it, dates it by season and information it on the fitting rail. Then construct a glance by hand on a drag-and-drop canvas, or ask a stylist for 3 — from garments the consumer already owns. 15 screens, gentle and darkish, and each one in every of them runs with no API key, no account and no community.
Demo Directions — the sign-in display is a mock and contacts no server, so any credentials are accepted. Use [email protected] / demo1234, or just faucet Sign up with each fields left empty. There may be nothing to register, no API key to enter and no community required: Demo mode ships switched on, so the closet, the canvas, the stylist and the insights all reply on the machine.
AI disclosure — two options on this app are produced by a third-party AI mannequin at run time: studying the tags off {a photograph} of a garment, and placing an outfit collectively out of the consumer’s personal closet. Nothing else within the app is AI-driven — each determine on the insights display is computed in plain Dart on the machine. The mannequin is bring-your-own-key: you or your customers provide a key for Google Gemini, OpenAI or Anthropic Claude, these companies invoice their very own utilization immediately, and no key or credit score is included with this merchandise. With no key in any respect the app nonetheless runs finish to finish, as a result of Demo mode assembles outfits on the machine from the closet in entrance of it. No asset on this bundle is AI-generated: the merchandise ships zero picture property — the pattern clothes, the colour swatches, the onboarding plates and the app mark are all drawn by a CustomPainter at run time, and the three bundled typefaces are SIL Open Font License 1.1 releases with the complete license textual content included. The one photos the app ever reveals as images are those its customers take themselves.
What’s within the bundle
- Full Flutter supply — 15 screens, gentle and darkish, GetX structure (View → Controller → Service → Mannequin, lazy bindings per route)
- An area cost-per-wear engine in plain Dart: wears, price per put on, dormant items and the closet’s palette, all computed on the machine from the wear and tear log
- Three AI suppliers as plain REST purchasers — Google Gemini (default), OpenAI, Anthropic Claude — bring-your-own-key, no vendor SDKs
- Demo mode, on out of the field: each display populated, each signature interplay working, with no key in any respect
- Offline HTML documentation — each display, a function walkthrough with two animated GIFs, supplier setup, connecting your personal storage, renaming the app, re-theming, challenge construction, a terminal glossary and troubleshooting
- 5 runtime dependencies in whole, and 0 picture property — each illustration within the UI is painted at run time
Who is that this for
Builders transport a wardrobe, closet or outfit app who need the arduous elements already determined: what the mannequin is requested for, what occurs when it solutions with one thing the consumer doesn’t personal, and the place the numbers on the insights display come from. The competing apps on this class are client merchandise with funding behind them; what’s lacking on this market is a template a developer can really get operating. This one runs on flutter run.
There isn’t any backend
No Firebase challenge, no admin panel, no server of yours to maintain up, and no account to your customers to make. Nothing on this merchandise telephones house — the one requests it could possibly make are those your consumer’s personal key pays for. The closet lives behind a service layer whose strategies already return Futures, so pointing it at Hive, Isar, SQLite or your personal API is a change in a single file and no display modifications. The only loudest grievance within the opinions of each competing app on this class is that the client couldn’t get the backend operating; this template doesn’t have one to get operating.
{Photograph} it and the tags fill in
Identify, class, shade, material, formality and season all come again from one image, in a single structured reply. Each one in every of them stays editable — a improper shade the consumer can not right is worse than no shade in any respect, and it’s the distinction between a device and a demo. Including a chunk by hand, with no {photograph} and no key, is identical kind with the identical fields.
Rails, not a grid
9 rails, one per class, every a hairline with the garments hanging underneath it. Faucet a rail head and it opens as a grid — which is the view you need if you end up looking for one particular factor, and the improper one for all the pieces else. A grid of forty tiles is a list report; a rail is how a wardrobe is definitely organized, and it’s the purpose this display reads as garments slightly than as rows.
Lay an outfit out by hand
Drag to maneuver, faucet to pick out, then resize or restack — convey to entrance, ship to again, pull one other piece from any rail, identify it and it’s a saved look. Placements are saved as fractions of the canvas, not pixels, so a glance organized on a telephone reopens accurately on a pill, and the identical placements draw the small collage on the look card with no second structure move. The entire outfit will be logged as worn in a single faucet, which is what feeds the cost-per-wear numbers.
Solely garments you personal
Two chips — the event and the climate, each chosen by the consumer, with no climate service to join — and a button. The immediate lists the closet with its ids and tells the mannequin to repeat them precisely; each id that comes again is seemed up earlier than something is drawn, ids that don’t resolve are dropped, and an outfit left with fewer than two actual items is thrown away entire slightly than proven with a gap in it. A stylist that implies a jacket you do not need is your entire failure mode of this class, and it’s why most “AI wardrobe” apps quietly turn into temper boards.
What each bit really price you
Worth divided by wears, counted on the machine from the wear and tear log — by no means saved on the garment, so it can not drift old-fashioned. It opens on the worst first, as a result of the coat you wore twice is the one no person goes in search of. Three trustworthy non-answers are in-built: a chunk with no value recorded, a chunk by no means worn, and a chunk purchased too lately to evaluate are all omitted slightly than given a made-up quantity. Alongside it: the closet’s palette as a stack of material swatches grouped by shade identify slightly than by hex — two navy items shot underneath totally different lights are nonetheless one navy — and the items that haven’t been worn in three months.
Three AI suppliers, or none in any respect
- Google Gemini — the shipped default. A secret is free and desires no bank card, so your customers are by no means despatched to a billing kind, and it does each jobs: it reads images and it writes textual content
- OpenAI and Anthropic Claude
- Demo mode — outfits are put collectively on the machine from the consumer’s personal closet, so the app solutions for garments they add themselves, with no key and no community. Not a folder of canned pattern outfits
Each supplier is a plain REST shopper behind a one-method interface that is aware of nothing about garments, which is why the identical three adapters serve each of this app’s very totally different requests — studying the tags off {a photograph}, and assembling an outfit out of a listing of ids. Including a fourth supplier is one class and one catalog entry, with no display modifications. Mannequin names expire, and this template is constructed for that: the mannequin id is an editable setting, a retired identify produces a plain-language error with a button straight to the sphere that fixes it, and the documentation has a bit on updating it.
15 screens, all wired up
Past the 4 signature screens: including a chunk by digital camera, gallery or hand; a garment’s personal web page with its value, its wears and its price per put on; search with three rows of composable aspects; a saved-looks grid and a glance element web page; onboarding; sign-in and sign-up; supplier settings; and a mock paywall. Each one in every of them themed twice, gentle and darkish, each full.
All the things, and nothing additional
- Full Flutter supply, GetX, analyzer clear — 63 assessments move, together with the cost-per-wear arithmetic, the three trustworthy non-answers, the per-vendor JSON schema transforms and the pattern closet’s personal integrity
- API keys saved within the Keychain on iOS and EncryptedSharedPreferences on Android — by no means in plain preferences
- All demo information in a single file: 38 clothes and 180 days of damage historical past with lifelike gaps in it, so the insights display has one thing true to say on first launch
- Zero picture property — the pattern clothes, the colour swatch stack, the onboarding plates and the app mark are all painted at run time, so there isn’t a paintings to license and nothing to maintain in sync. Clothes your customers {photograph} are proven as images
- Typefaces are bundled, not downloaded — the UI renders identically offline and on a chilly first launch
- A particular fashion-desk design system: heat charcoal on cool linen with a single saturated blue used just for the issues you are able to do, Syne + Onest typography, and a tailor-card form — the top-right nook reduce off the way in which a tailor notches a sample piece, with a dashed topstitch drawn in the identical move because the fill. Each token is centralized, so the entire app re-skins from one shade file
Get it operating in 5 minutes
- Unzip, then
flutter pub get flutter run— Demo mode is already on, so the app is totally populated- Elective: paste a supplier key in Settings to {photograph} an actual garment and have a stay mannequin tag it
Necessities
To run and construct the supply code you want the Flutter SDK (model 3.29 or newer) and Dart 3.7 or newer, along with a Flutter-ready IDE — Android Studio or Visible Studio Code with the Flutter & Dart plugins. Runs on each Android and iOS from a single Flutter codebase (Materials 3, null-safe).
Third-party prices and trustworthy limits
Capsora runs fully on the machine, however the two AI options — studying the tags off {a photograph}, and placing an outfit collectively — are produced by a third-party AI supplier that you just or your customers provide a key for. Connecting it to a stay supplier reminiscent of Google Gemini, OpenAI or Anthropic will contain these third-party companies, which will incur their very own separate prices and aren’t included on this merchandise. Pictures a consumer provides are downscaled on the machine and despatched solely to the supplier whose key that consumer equipped; nothing is uploaded wherever else. The tags a mannequin returns are its studying of an image, not a measurement — which is why each one in every of them stays editable.
What isn’t included
- No backend and no account system. Signal-in is a mock that contacts no server — the seam is drawn the place your API name goes, and the documentation walks via it
- The closet is held in reminiscence. It’s seeded with 38 pattern clothes and resets when the app restarts; wiring a persistent retailer is a documented one-file change and is left to you. The one factor written to disk is the consumer’s API key
- No digital try-on, no background removing and no picture era of any form. This template reads images and writes textual content
- No climate service. The climate the stylist works from is a chip the consumer picks, not a stay forecast
- The subscription display is a mock — no fee code, receipt validation or retailer product is included
- No push or native notifications / reminders — nothing within the bundle schedules something
- No AI supplier key or credit score is included, and no net or desktop construct is shipped or supported
Help and updates
Questions in regards to the merchandise undergo the feedback tab on this web page, which is the place I learn them. Help covers the code because it ships: getting the challenge to construct, the documented setup steps, supplier configuration, and defects in what’s included. It doesn’t cowl your personal modifications, third-party supplier outages or pricing modifications, or constructing new options to order.
Changelog
v1.0 — preliminary launch.
