diff --git a/README.ar.md b/README.ar.md
new file mode 100644
index 0000000..f2075a8
--- /dev/null
+++ b/README.ar.md
@@ -0,0 +1,118 @@
+
+
+
+
+# GLB_FACTORY
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+**استوديو تفاعلي لتحويل الصور الشخصية إلى أفاتار ثلاثي الأبعاد.** ارفعي صورة شخصية، ودعي
+التطبيق يقرأ الوجه ولون البشرة والشعر والملابس، ثم صدّري نموذج **GLB** جاهزًا بالكامل
+لمحركات الألعاب، من دون أي خبرة في النمذجة ثلاثية الأبعاد.
+
+التطبيق المباشر: [dacameragirl.github.io/GLB_FACTORY](https://dacameragirl.github.io/GLB_FACTORY/)
+
+---
+
+## أبرز المزايا
+
+| الميزة | ماذا تفعل |
+|---|---|
+| **صورة ← أفاتار** | ارفعي صورة شخصية واحدة واحصلي على شخصية ثلاثية الأبعاد جاهزة للألعاب |
+| **استخراج الألوان اعتمادًا على الوجه** | يكتشف لون البشرة والشعر والملابس مباشرة من الصورة |
+| **اقتراح تسريحة الشعر** | يقترح نوع تسريحة مناسبًا بناءً على الصورة الأصلية |
+| **معاينة ثلاثية الأبعاد مباشرة** | دوّري وكبّري وافحصي الأفاتار الناتج في عارض Three.js قبل التصدير |
+| **تصدير GLB بنقرة واحدة** | نزّلي ملف `.glb` قياسي جاهزًا لمحركات الألعاب وعارضات النماذج ثلاثية الأبعاد |
+| **يعمل بوجود خادم أو من دونه** | تحليل كامل عبر Gemini عند الاستضافة، وتحوّل تلقائي إلى تحليل بالمتصفح في الوضع الساكن |
+
+## بنية ذات وضعين
+
+صُمم GLB_FACTORY للعمل بطريقتين مختلفتين حسب مكان نشره، ويختار الطريقة المناسبة تلقائيًا:
+
+1. **الوضع المدعوم بالذكاء الاصطناعي (استضافة Node/Express)** — في بيئة متكاملة مثل
+ التطوير المحلي أو حاوية سحابية، يتواصل التطبيق مع وسيط خلفي متصل بـ **Gemini 3.5 Flash**.
+ يحدد Gemini موقع الوجه تلقائيًا ويستخرج لون البشرة ولون الشعر ولون الملابس وتسريحة
+ مقترحة بدقة بصرية عالية.
+
+2. **وضع الاحتياط الساكن (GitHub Pages)** — عند عدم توفر خادم خلفي، يكتشف التطبيق البيئة
+ الساكنة وينتقل إلى **تحليل الوجه من جهة المتصفح**: يقرأ عيّن كانفاس HTML5 خفيف بيانات
+ بكسل الصورة مباشرة داخل المتصفح، ويستخرج نفس ألوان البشرة والشعر والملابس من دون أي
+ طلبات شبكة.
+
+نفس الواجهة، ونفس نتيجة GLB، لكن بمحركين مختلفين خلف الكواليس، بحسب ما يمكن لبيئة النشر
+تشغيله فعليًا.
+
+---
+
+## البدء السريع
+
+```bash
+npm install
+npm run dev
+```
+
+افتحي [http://localhost:3000](http://localhost:3000) في المتصفح.
+
+لتفعيل التحليل بالذكاء الاصطناعي محليًا، أضيفي مفتاح Gemini:
+
+```env
+# .env.local
+GEMINI_API_KEY=your_gemini_api_key_here
+```
+
+من دون مفتاح، يظل التطبيق يعمل، ويستخدم فقط محلل الاحتياط من جهة المتصفح.
+
+## النشر على GitHub Pages
+
+يتضمن المستودع ملف `.github/workflows/deploy.yml` الذي يبني وينشر التطبيق الساكن مع كل
+دفع (push) إلى فرع `main`.
+
+1. اذهبي إلى المستودع على GitHub، ثم **Settings**.
+2. ضمن **Code and automation → Pages**، اضبطي **Source** على **GitHub Actions**.
+3. ادفعي إلى `main` وتابعي عملية البناء من تبويب **Actions**.
+
+---
+
+## التقنيات المستخدمة
+
+| الطبقة | التقنية |
+|---|---|
+| العرض ثلاثي الأبعاد | **Three.js** — عرض WebGL وبناء شبكي إجرائي للأفاتار |
+| الواجهة الأمامية | **React 19** + **Vite 6** — بيئة تشغيل وبناء التطبيق أحادي الصفحة |
+| التنسيق | **Tailwind CSS v4** |
+| الأيقونات | **Lucide React** |
+| الخلفية | **Express** + **Google GenAI SDK** — وسيط لواجهة برمجة Gemini |
+
+## المساهمون
+
+- Angela — توجيه المنتج، الاختبار
+- Claude — التنفيذ وسير عمل GitHub
+
+## إشعار قانوني
+
+تُعالج الصور المرفوعة فقط بغرض توليد أفاتار ثلاثي الأبعاد. في الوضع المدعوم بالذكاء
+الاصطناعي، تُرسل بيانات الصورة إلى واجهة Gemini وفق شروط Google؛ وفي وضع الاحتياط الساكن،
+يتم التحليل بالكامل داخل المتصفح ولا تغادر أي بيانات الجهاز.
diff --git a/README.de.md b/README.de.md
new file mode 100644
index 0000000..b90e15e
--- /dev/null
+++ b/README.de.md
@@ -0,0 +1,121 @@
+
+
+
+
+# GLB_FACTORY
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+**Ein interaktives 3D-Foto-zu-Avatar-Studio.** Lade ein Porträt hoch, lass die App Gesicht,
+Hautton, Haar- und Kleidungsfarbe erkennen und exportiere ein voll einsatzbereites
+**GLB**-Modell für Spiele-Engines, ganz ohne 3D-Modellierungserfahrung.
+
+Live-App: [dacameragirl.github.io/GLB_FACTORY](https://dacameragirl.github.io/GLB_FACTORY/)
+
+---
+
+## Wichtigste Funktionen
+
+| Funktion | Was sie tut |
+|---|---|
+| **Foto → Avatar** | Ein einzelnes Porträt hochladen und einen spielfertigen 3D-Charakter erhalten |
+| **Gesichtsbasierte Farberkennung** | Erkennt Hautton, Haar- und Kleidungsfarbe direkt aus dem Foto |
+| **Frisurenempfehlung** | Schlägt einen passenden Frisurtyp basierend auf dem Quellbild vor |
+| **Live-3D-Vorschau** | Den generierten Avatar vor dem Export in einer Three.js-Ansicht drehen und zoomen |
+| **GLB-Export mit einem Klick** | Eine standardmäßige `.glb`-Datei für Spiele-Engines und 3D-Viewer herunterladen |
+| **Funktioniert mit oder ohne Server** | Volle Gemini-Analyse bei Hosting, automatischer Canvas-Fallback bei statischem Betrieb |
+
+## Zwei-Modi-Architektur
+
+GLB_FACTORY läuft je nach Deployment-Umgebung auf zwei unterschiedliche Arten und wählt die
+passende automatisch:
+
+1. **KI-gestützter Modus (Node/Express-Hosting)** — In einer vollständigen Umgebung wie
+ lokaler Entwicklung oder einem Cloud-Container spricht die App mit einem Backend-Proxy,
+ der an **Gemini 3.5 Flash** angebunden ist. Gemini lokalisiert automatisch das Gesicht
+ und liest Hautton, Haarfarbe, Kleidungsfarbe sowie eine empfohlene Frisur mit hoher
+ visueller Genauigkeit aus.
+
+2. **Statischer Fallback-Modus (GitHub Pages)** — Ohne verfügbares Backend erkennt die App
+ die statische Umgebung und wechselt zur **clientseitigen Gesichtsanalyse**: Ein
+ leichtgewichtiger HTML5-Canvas-Sampler liest die Pixel des Porträts direkt im Browser aus
+ und extrahiert dieselben Farben, ganz ohne Netzwerkanfragen.
+
+Gleiche Oberfläche, gleiches GLB-Ergebnis, zwei unterschiedliche Engines im Hintergrund, je
+nachdem, was die Deployment-Umgebung tatsächlich ausführen kann.
+
+---
+
+## Schnellstart
+
+```bash
+npm install
+npm run dev
+```
+
+Öffne [http://localhost:3000](http://localhost:3000) im Browser.
+
+Für die KI-gestützte Analyse lokal einen Gemini-Schlüssel hinzufügen:
+
+```env
+# .env.local
+GEMINI_API_KEY=your_gemini_api_key_here
+```
+
+Ohne Schlüssel läuft die App trotzdem, sie nutzt dann einfach den clientseitigen Fallback.
+
+## Deployment auf GitHub Pages
+
+Das Repository enthält `.github/workflows/deploy.yml`, das die statische App bei jedem
+Push auf `main` baut und veröffentlicht.
+
+1. Zum Repository auf GitHub gehen, dann **Settings**.
+2. Unter **Code and automation → Pages** bei **Source** **GitHub Actions** auswählen.
+3. Auf `main` pushen und den Build im Tab **Actions** verfolgen.
+
+---
+
+## Verwendete Technologien
+
+| Ebene | Stack |
+|---|---|
+| 3D-Rendering | **Three.js** — WebGL-Rendering und prozedurale Avatar-Mesh-Erstellung |
+| Frontend | **React 19** + **Vite 6** — SPA-Runtime und Build |
+| Styling | **Tailwind CSS v4** |
+| Icons | **Lucide React** |
+| Backend | **Express** + **Google GenAI SDK** — Gemini-API-Proxy |
+
+## Mitwirkende
+
+- Angela — Produktrichtung, Tests
+- Claude — Implementierung und GitHub-Workflow
+
+## Rechtliches
+
+Hochgeladene Fotos werden ausschließlich zur Erstellung eines 3D-Avatars verarbeitet. Im
+KI-Modus werden Bilddaten gemäß den Bedingungen von Google an die Gemini-API gesendet; im
+statischen Fallback-Modus läuft die Analyse vollständig im Browser, und nichts verlässt das
+Gerät.
diff --git a/README.es.md b/README.es.md
new file mode 100644
index 0000000..0040b58
--- /dev/null
+++ b/README.es.md
@@ -0,0 +1,120 @@
+
+
+
+
+# GLB_FACTORY
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+**Un estudio interactivo de foto a avatar 3D.** Sube un retrato, deja que la app detecte el
+rostro, el tono de piel, el cabello y los colores de la ropa, y exporta un modelo **GLB**
+completamente listo para usar en videojuegos, sin necesidad de experiencia en modelado 3D.
+
+App en vivo: [dacameragirl.github.io/GLB_FACTORY](https://dacameragirl.github.io/GLB_FACTORY/)
+
+---
+
+## Funciones principales
+
+| Función | Qué hace |
+|---|---|
+| **Foto → avatar** | Sube un solo retrato y obtén un personaje 3D estilizado, listo para juegos |
+| **Extracción de color a partir del rostro** | Detecta el tono de piel, el color del cabello y de la ropa desde la foto |
+| **Recomendación de peinado** | Sugiere un estilo de cabello acorde a lo que ve en la imagen original |
+| **Vista previa 3D en vivo** | Rota, acerca e inspecciona el avatar generado en un visor Three.js antes de exportar |
+| **Exportación GLB con un clic** | Descarga un archivo `.glb` estándar listo para motores de juego y visores 3D |
+| **Funciona con o sin servidor** | Análisis completo con Gemini si hay servidor, o alternativa por canvas si es estático |
+
+## Arquitectura de doble modo
+
+GLB_FACTORY está diseñado para funcionar de dos formas distintas según dónde se despliegue,
+y elige la correcta automáticamente:
+
+1. **Modo con IA (alojamiento Node/Express)** — En un entorno completo como desarrollo
+ local o un contenedor en la nube, la app habla con un proxy de backend conectado a
+ **Gemini 3.5 Flash**. Gemini ubica el rostro automáticamente y extrae el tono de piel,
+ color de cabello, color de ropa y un estilo de peinado recomendado con alta precisión.
+
+2. **Modo estático de respaldo (GitHub Pages)** — Sin backend disponible, la app detecta
+ el entorno estático y cambia a **análisis facial del lado del cliente**: un ligero
+ muestreador de canvas HTML5 lee los píxeles del retrato directamente en el navegador y
+ extrae los mismos colores de piel, cabello y ropa sin ninguna solicitud de red.
+
+Misma interfaz, mismo resultado GLB, dos motores distintos por debajo, el que el entorno de
+despliegue pueda ejecutar realmente.
+
+---
+
+## Inicio rápido
+
+```bash
+npm install
+npm run dev
+```
+
+Abre [http://localhost:3000](http://localhost:3000) en tu navegador.
+
+Para habilitar el análisis con IA en local, agrega una clave de Gemini:
+
+```env
+# .env.local
+GEMINI_API_KEY=your_gemini_api_key_here
+```
+
+Sin la clave, la app sigue funcionando: simplemente usa el analizador de respaldo del lado
+del cliente.
+
+## Despliegue en GitHub Pages
+
+El repositorio incluye `.github/workflows/deploy.yml`, que construye y publica la app
+estática en cada push a `main`.
+
+1. Ve al repositorio en GitHub, luego **Settings**.
+2. En **Code and automation → Pages**, elige **GitHub Actions** como **Source**.
+3. Haz push a `main` y observa la compilación en la pestaña **Actions**.
+
+---
+
+## Tecnologías utilizadas
+
+| Capa | Stack |
+|---|---|
+| Renderizado 3D | **Three.js** — renderizado WebGL y construcción procedural de la malla del avatar |
+| Frontend | **React 19** + **Vite 6** — runtime de la SPA y compilación |
+| Estilos | **Tailwind CSS v4** |
+| Iconos | **Lucide React** |
+| Backend | **Express** + **Google GenAI SDK** — proxy de la API de Gemini |
+
+## Colaboradores
+
+- Angela — dirección de producto, pruebas
+- Claude — implementación y flujo de trabajo en GitHub
+
+## Aviso legal
+
+Las fotos subidas se procesan con el único fin de generar un avatar 3D. En modo con IA, los
+datos de la imagen se envían a la API de Gemini bajo los términos de Google; en modo estático
+de respaldo, el análisis ocurre completamente en el navegador y nada sale del dispositivo.
diff --git a/README.fr.md b/README.fr.md
new file mode 100644
index 0000000..e3316a8
--- /dev/null
+++ b/README.fr.md
@@ -0,0 +1,121 @@
+
+
+
+
+# GLB_FACTORY
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+**Un studio interactif de photo vers avatar 3D.** Chargez un portrait, laissez l'application
+lire le visage, le teint, la couleur des cheveux et de la tenue, puis exportez un modèle
+**GLB** complet, prêt pour le jeu vidéo, sans aucune expérience en modélisation 3D.
+
+Application en ligne : [dacameragirl.github.io/GLB_FACTORY](https://dacameragirl.github.io/GLB_FACTORY/)
+
+---
+
+## Fonctionnalités principales
+
+| Fonctionnalité | Ce qu'elle fait |
+|---|---|
+| **Photo → avatar** | Chargez un simple portrait et obtenez un personnage 3D stylisé, prêt pour le jeu |
+| **Extraction des couleurs à partir du visage** | Détecte le teint, la couleur des cheveux et de la tenue à partir de la photo |
+| **Recommandation de coiffure** | Suggère un style de coiffure adapté à ce qu'elle détecte sur l'image source |
+| **Aperçu 3D en direct** | Faites pivoter, zoomez et inspectez l'avatar généré dans une vue Three.js avant export |
+| **Export GLB en un clic** | Téléchargez un fichier `.glb` standard prêt pour moteurs de jeu et visionneuses 3D |
+| **Fonctionne avec ou sans serveur** | Analyse complète via Gemini si hébergé, repli automatique par canvas si statique |
+
+## Architecture à double mode
+
+GLB_FACTORY est conçu pour fonctionner de deux façons différentes selon l'endroit où il est
+déployé, et choisit automatiquement la bonne :
+
+1. **Mode IA (hébergement Node/Express)** — Sur un environnement complet comme le
+ développement local ou un conteneur cloud, l'application communique avec un proxy backend
+ relié à **Gemini 3.5 Flash**. Gemini localise automatiquement le visage et en extrait le
+ teint, la couleur des cheveux, la couleur de la tenue et une coiffure recommandée avec une
+ grande précision visuelle.
+
+2. **Mode de repli statique (GitHub Pages)** — Sans backend disponible, l'application
+ détecte l'environnement statique et bascule vers une **analyse faciale côté client** :
+ un échantillonneur canvas HTML5 léger lit les pixels du portrait directement dans le
+ navigateur et extrait les mêmes couleurs, sans aucune requête réseau.
+
+Même interface, même résultat GLB, deux moteurs différents en coulisses, selon ce que
+l'environnement de déploiement peut réellement exécuter.
+
+---
+
+## Démarrage rapide
+
+```bash
+npm install
+npm run dev
+```
+
+Ouvrez [http://localhost:3000](http://localhost:3000) dans votre navigateur.
+
+Pour activer l'analyse IA en local, ajoutez une clé Gemini :
+
+```env
+# .env.local
+GEMINI_API_KEY=your_gemini_api_key_here
+```
+
+Sans clé, l'application fonctionne quand même : elle utilise simplement l'analyseur de
+repli côté client.
+
+## Déploiement sur GitHub Pages
+
+Le dépôt inclut `.github/workflows/deploy.yml`, qui construit et publie l'application
+statique à chaque push sur `main`.
+
+1. Allez sur le dépôt GitHub, puis **Settings**.
+2. Sous **Code and automation → Pages**, choisissez **GitHub Actions** comme **Source**.
+3. Poussez sur `main` et suivez la construction dans l'onglet **Actions**.
+
+---
+
+## Technologies utilisées
+
+| Couche | Stack |
+|---|---|
+| Rendu 3D | **Three.js** — rendu WebGL et construction procédurale du maillage de l'avatar |
+| Frontend | **React 19** + **Vite 6** — runtime et build de la SPA |
+| Style | **Tailwind CSS v4** |
+| Icônes | **Lucide React** |
+| Backend | **Express** + **Google GenAI SDK** — proxy vers l'API Gemini |
+
+## Contributeurs
+
+- Angela — direction produit, tests
+- Claude — implémentation et workflow GitHub
+
+## Mentions légales
+
+Les photos envoyées sont traitées dans le seul but de générer un avatar 3D. En mode IA, les
+données de l'image sont envoyées à l'API Gemini selon les conditions de Google ; en mode de
+repli statique, l'analyse se fait entièrement dans le navigateur et rien ne quitte l'appareil.
diff --git a/README.it.md b/README.it.md
new file mode 100644
index 0000000..ba599eb
--- /dev/null
+++ b/README.it.md
@@ -0,0 +1,122 @@
+
+
+
+
+# GLB_FACTORY
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+**Uno studio interattivo per trasformare foto in avatar 3D.** Carica un ritratto, lascia che
+l'app legga viso, incarnato, colore dei capelli e dell'abbigliamento, ed esporta un modello
+**GLB** pronto per motori di gioco, senza alcuna esperienza di modellazione 3D.
+
+App live: [dacameragirl.github.io/GLB_FACTORY](https://dacameragirl.github.io/GLB_FACTORY/)
+
+---
+
+## Funzionalità principali
+
+| Funzionalità | Cosa fa |
+|---|---|
+| **Foto → avatar** | Carica un singolo ritratto e ottieni un personaggio 3D stilizzato, pronto per i giochi |
+| **Estrazione colore basata sul viso** | Rileva incarnato, colore dei capelli e dell'abbigliamento direttamente dalla foto |
+| **Consiglio di acconciatura** | Suggerisce uno stile di capelli adatto in base all'immagine originale |
+| **Anteprima 3D live** | Ruota, ingrandisci e ispeziona l'avatar generato in una vista Three.js prima di esportare |
+| **Esportazione GLB in un clic** | Scarica un file `.glb` standard pronto per motori di gioco e visualizzatori 3D |
+| **Funziona con o senza server** | Analisi completa con Gemini se ospitato, fallback automatico via canvas se statico |
+
+## Architettura a doppia modalità
+
+GLB_FACTORY è progettato per funzionare in due modi diversi a seconda di dove viene
+distribuito, e sceglie automaticamente quello giusto:
+
+1. **Modalità con IA (hosting Node/Express)** — In un ambiente completo come lo sviluppo
+ locale o un container cloud, l'app comunica con un proxy backend collegato a
+ **Gemini 3.5 Flash**. Gemini localizza automaticamente il viso ed estrae incarnato,
+ colore dei capelli, colore dell'abbigliamento e un'acconciatura consigliata con alta
+ precisione visiva.
+
+2. **Modalità di fallback statico (GitHub Pages)** — Senza un backend disponibile, l'app
+ rileva l'ambiente statico e passa all'**analisi del viso lato client**: un leggero
+ campionatore canvas HTML5 legge i pixel del ritratto direttamente nel browser ed estrae
+ gli stessi colori, senza alcuna richiesta di rete.
+
+Stessa interfaccia, stesso risultato GLB, due motori diversi dietro le quinte, a seconda di
+cosa l'ambiente di distribuzione può effettivamente eseguire.
+
+---
+
+## Avvio rapido
+
+```bash
+npm install
+npm run dev
+```
+
+Apri [http://localhost:3000](http://localhost:3000) nel browser.
+
+Per abilitare l'analisi con IA in locale, aggiungi una chiave Gemini:
+
+```env
+# .env.local
+GEMINI_API_KEY=your_gemini_api_key_here
+```
+
+Senza chiave, l'app funziona comunque: usa semplicemente l'analizzatore di fallback lato
+client.
+
+## Distribuzione su GitHub Pages
+
+Il repository include `.github/workflows/deploy.yml`, che costruisce e pubblica l'app
+statica a ogni push su `main`.
+
+1. Vai al repository su GitHub, poi **Settings**.
+2. In **Code and automation → Pages**, imposta **Source** su **GitHub Actions**.
+3. Fai push su `main` e segui la build nella scheda **Actions**.
+
+---
+
+## Tecnologie utilizzate
+
+| Livello | Stack |
+|---|---|
+| Rendering 3D | **Three.js** — rendering WebGL e costruzione procedurale della mesh dell'avatar |
+| Frontend | **React 19** + **Vite 6** — runtime e build della SPA |
+| Stile | **Tailwind CSS v4** |
+| Icone | **Lucide React** |
+| Backend | **Express** + **Google GenAI SDK** — proxy per l'API Gemini |
+
+## Collaboratori
+
+- Angela — direzione di prodotto, test
+- Claude — implementazione e workflow GitHub
+
+## Note legali
+
+Le foto caricate vengono elaborate al solo scopo di generare un avatar 3D. In modalità con
+IA, i dati dell'immagine vengono inviati all'API Gemini secondo i termini di Google; in
+modalità di fallback statico, l'analisi avviene interamente nel browser e nessun dato lascia
+il dispositivo.
diff --git a/README.ja.md b/README.ja.md
new file mode 100644
index 0000000..1b3827c
--- /dev/null
+++ b/README.ja.md
@@ -0,0 +1,118 @@
+
+
+**인터랙티브한 3D 사진-아바타 스튜디오.** 인물 사진을 업로드하면 앱이 얼굴, 피부톤, 머리색,
+옷 색상을 읽어들여 3D 모델링 경험이 없어도 게임 엔진에 바로 쓸 수 있는 완전한 **GLB** 모델을
+내보냅니다.
+
+라이브 앱: [dacameragirl.github.io/GLB_FACTORY](https://dacameragirl.github.io/GLB_FACTORY/)
+
+---
+
+## 주요 기능
+
+| 기능 | 설명 |
+|---|---|
+| **사진 → 아바타** | 인물 사진 한 장만 업로드하면 게임에 바로 쓸 수 있는 3D 캐릭터 생성 |
+| **얼굴 기반 색상 추출** | 사진에서 피부톤, 머리색, 옷 색상을 직접 감지 |
+| **헤어스타일 추천** | 원본 이미지에 맞는 헤어스타일 제안 |
+| **실시간 3D 미리보기** | 내보내기 전에 Three.js 뷰포트에서 회전, 확대하며 확인 |
+| **원클릭 GLB 내보내기** | 게임 엔진과 3D 뷰어에서 바로 쓸 수 있는 표준 `.glb` 파일 다운로드 |
+| **서버 유무와 관계없이 동작** | 호스팅 환경에서는 Gemini로 전체 분석, 정적 환경에서는 캔버스 분석으로 자동 전환 |
+
+## 듀얼 모드 아키텍처
+
+GLB_FACTORY는 배포 환경에 따라 두 가지 방식으로 동작하며, 알맞은 방식을 자동으로 선택합니다.
+
+1. **AI 기반 모드 (Node/Express 호스팅)** — 로컬 개발이나 클라우드 컨테이너처럼 완전한
+ 환경에서는 **Gemini 3.5 Flash**에 연결된 백엔드 프록시와 통신합니다. Gemini가 얼굴 위치를
+ 자동으로 찾아내고, 피부톤·머리색·옷 색상과 추천 헤어스타일을 높은 정확도로 추출합니다.
+
+2. **정적 폴백 모드 (GitHub Pages)** — 백엔드를 사용할 수 없을 때, 앱은 정적 환경을 감지해
+ **클라이언트 측 얼굴 분석**으로 전환합니다. 가벼운 HTML5 캔버스 샘플러가 브라우저 안에서
+ 직접 인물 사진의 픽셀 데이터를 읽어, 네트워크 요청 없이 동일한 피부·머리·옷 색상을 추출합니다.
+
+동일한 UI, 동일한 GLB 결과물, 배포 환경이 실제로 실행할 수 있는 쪽에 따라 내부 엔진만 다르게
+동작합니다.
+
+---
+
+## 빠른 시작
+
+```bash
+npm install
+npm run dev
+```
+
+브라우저에서 [http://localhost:3000](http://localhost:3000) 을 여세요.
+
+로컬에서 AI 분석을 활성화하려면 Gemini 키를 추가하세요.
+
+```env
+# .env.local
+GEMINI_API_KEY=your_gemini_api_key_here
+```
+
+키가 없어도 앱은 정상 동작하며, 클라이언트 측 폴백 분석기를 사용합니다.
+
+## GitHub Pages 배포
+
+이 저장소에는 `.github/workflows/deploy.yml` 이 포함되어 있어, `main` 에 push할 때마다
+정적 앱을 빌드하고 배포합니다.
+
+1. GitHub 저장소로 이동한 다음 **Settings** 를 엽니다.
+2. **Code and automation → Pages** 에서 **Source** 를 **GitHub Actions** 로 설정합니다.
+3. `main` 에 push하고 **Actions** 탭에서 빌드 진행 상황을 확인합니다.
+
+---
+
+## 사용 기술
+
+| 레이어 | 스택 |
+|---|---|
+| 3D 렌더링 | **Three.js** — WebGL 렌더링과 아바타 메시의 절차적 생성 |
+| 프론트엔드 | **React 19** + **Vite 6** — SPA 런타임과 빌드 |
+| 스타일링 | **Tailwind CSS v4** |
+| 아이콘 | **Lucide React** |
+| 백엔드 | **Express** + **Google GenAI SDK** — Gemini API 프록시 |
+
+## 기여자
+
+- Angela — 제품 방향, 테스트
+- Claude — 구현 및 GitHub 워크플로
+
+## 법적 고지
+
+업로드된 사진은 오직 3D 아바타 생성 목적으로만 처리됩니다. AI 기반 모드에서는 이미지 데이터가
+Google의 약관에 따라 Gemini API로 전송됩니다. 정적 폴백 모드에서는 분석이 전적으로 브라우저
+안에서 이루어지며, 어떤 데이터도 기기 밖으로 나가지 않습니다.
diff --git a/README.md b/README.md
index c921054..987df77 100644
--- a/README.md
+++ b/README.md
@@ -1,69 +1,119 @@
-# GLB_FACTORY 🎨🤖
+
+
+
+
+# GLB_FACTORY
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+**An interactive 3D Photo-to-Avatar Studio.** Upload a portrait, let the app read the face,
+skin tone, hair, and outfit colors, then export a fully rigged, game-ready **GLB** model —
+no 3D modeling experience required.
+
+Live app: [dacameragirl.github.io/GLB_FACTORY](https://dacameragirl.github.io/GLB_FACTORY/)
-Welcome to **GLB_FACTORY**! This application is an interactive **3D Photo-to-Avatar Studio**. It allows users to upload a portrait photo and automatically generate, customize, and export fully functional 3D blocky models in standard **GLB** format, ready for game engines or 3D viewports.
+---
-View the application live on GitHub Pages: [dacameragirl.github.io/GLB_FACTORY/](https://dacameragirl.github.io/GLB_FACTORY/)
+## Product highlights
----
+| Feature | What it does |
+|---|---|
+| **Photo → avatar** | Upload a single portrait and get a blocky, game-ready 3D character back |
+| **Face-aware color extraction** | Detects skin tone, hair color, and clothing color straight from the photo |
+| **Hairstyle recommendation** | Suggests a fitting hairstyle type based on what it sees in the source image |
+| **Live 3D preview** | Rotate, zoom, and inspect the generated avatar in a Three.js viewport before exporting |
+| **One-click GLB export** | Download a standard `.glb` file ready for game engines and 3D viewers |
+| **Works with or without a server** | Full Gemini-powered analysis when hosted, automatic canvas-based fallback when static |
+
+## Dual-mode architecture
-## 🚀 Dual-Mode Architecture
+GLB_FACTORY is built to run two different ways depending on where it's deployed, and it
+picks the right one automatically:
-This application is built with a highly resilient **hybrid architecture**:
+1. **AI-powered mode (Node/Express hosting)** — On a full-stack environment like local
+ dev or a container host, the app talks to a backend proxy wired to **Gemini 3.5 Flash**.
+ Gemini locates the face bounding box and reads skin tone, hair color, clothing color, and
+ a recommended hairstyle with high visual precision.
-1. **AI-Powered Mode (Cloud Hosting)**:
- - When running on a full-stack Node/Express container environment (like local development or Cloud Run), the app communicates with a backend proxy connected to the **Gemini 3.5 Flash** API.
- - Gemini automatically locates the face bounding box, extracts skin tones, hair colors, clothing colors, and recommends fitting hairstyle types with high visual precision.
+2. **Static fallback mode (GitHub Pages)** — With no backend available, the app detects the
+ static environment and switches to **client-side face analysis**: a lightweight HTML5
+ canvas sampler reads the portrait's pixel data directly in the browser and extracts the
+ same skin, hair, and clothing colors with zero network requests.
-2. **Static Fallback Mode (GitHub Pages)**:
- - When deployed statically on **GitHub Pages**, where no custom backend server runs, the app **automatically detects the environment** and switches to **Client-Side Face Analysis**.
- - It utilizes a lightweight HTML5 canvas sampler to analyze the pixel data of the loaded portrait, extracting the representative skin, hair, and clothing colors directly in the browser with zero external network requests!
+Same UI, same GLB output, two different engines under the hood, whichever one the
+deployment target can actually run.
---
-## ⚡ GitHub Pages Deployment
+## Quick start
-The repository is equipped with an automated GitHub Actions workflow (`.github/workflows/deploy.yml`) that builds and publishes the application dynamically on every commit to the `main` branch.
+```bash
+npm install
+npm run dev
+```
-### How to Activate GitHub Pages in Your Repository:
-1. Go to your repository on GitHub: `https://github.com/DaCameraGirl/GLB_FACTORY`.
-2. Click on the **Settings** tab.
-3. In the left sidebar, navigate to **Pages** under the *Code and automation* section.
-4. Under **Build and deployment**:
- - For **Source**, select **GitHub Actions** from the dropdown.
-5. Once selected, your automated workflow will automatically build and publish the static app.
-6. The deployment progress can be monitored under the **Actions** tab.
+Open [http://localhost:3000](http://localhost:3000) in your browser.
----
+To enable AI-powered analysis locally, add a Gemini key:
-## 🛠️ Local Development
+```env
+# .env.local
+GEMINI_API_KEY=your_gemini_api_key_here
+```
-### Prerequisites
-- Node.js (v18+)
-- npm
+Without a key, the app still runs, it just uses the client-side fallback analyzer.
-### Installation & Run
-1. Install dependencies:
- ```bash
- npm install
- ```
+## Deploying to GitHub Pages
-2. Configure environment variables (optional for Gemini server features):
- Create a `.env.local` or `.env` file in the root:
- ```env
- GEMINI_API_KEY=your_gemini_api_key_here
- ```
+The repo ships with `.github/workflows/deploy.yml`, which builds and publishes the static
+app on every push to `main`.
-3. Start the developmental server:
- ```bash
- npm run dev
- ```
- Open [http://localhost:3000](http://localhost:3000) in your browser.
+1. Go to the repo on GitHub, then **Settings**.
+2. Under **Code and automation → Pages**, set **Source** to **GitHub Actions**.
+3. Push to `main` and watch the build under the **Actions** tab.
---
-## 🏗️ Technologies Used
-- **Three.js** (WebGL 3D Rendering & Procedural Avatar Mesh Construction)
-- **React 19** + **Vite 6** (Modern SPA runtime & high-speed builder)
-- **Tailwind CSS v4** (Modern utility-first responsive layout styling)
-- **Lucide React** (Premium, lightweight icon pairings)
-- **Express + Google GenAI SDK** (Lightweight backend proxy handling Gemini API orchestration)
+## Technologies used
+
+| Layer | Stack |
+|---|---|
+| 3D rendering | **Three.js** — WebGL rendering and procedural avatar mesh construction |
+| Frontend | **React 19** + **Vite 6** — SPA runtime and build |
+| Styling | **Tailwind CSS v4** |
+| Icons | **Lucide React** |
+| Backend | **Express** + **Google GenAI SDK** — Gemini API proxy |
+
+## Contributors
+
+- Angela — product direction, testing
+- Claude — implementation and GitHub workflow
+
+## Legal
+
+Uploaded photos are processed for the sole purpose of generating a 3D avatar. In AI-powered
+mode, image data is sent to the Gemini API under Google's terms; in static fallback mode,
+analysis happens entirely in the browser and nothing leaves the device.
diff --git a/README.pt-BR.md b/README.pt-BR.md
new file mode 100644
index 0000000..c777970
--- /dev/null
+++ b/README.pt-BR.md
@@ -0,0 +1,120 @@
+
+
+
+
+# GLB_FACTORY
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+**Um estúdio interativo de foto para avatar 3D.** Envie um retrato, deixe o app ler o rosto,
+o tom de pele, o cabelo e as cores da roupa, e exporte um modelo **GLB** pronto para motores
+de jogos, sem precisar de experiência em modelagem 3D.
+
+App ao vivo: [dacameragirl.github.io/GLB_FACTORY](https://dacameragirl.github.io/GLB_FACTORY/)
+
+---
+
+## Principais funcionalidades
+
+| Funcionalidade | O que faz |
+|---|---|
+| **Foto → avatar** | Envie um único retrato e receba um personagem 3D estilizado, pronto para jogos |
+| **Extração de cor a partir do rosto** | Detecta tom de pele, cor do cabelo e da roupa direto da foto |
+| **Recomendação de penteado** | Sugere um estilo de cabelo compatível com o que vê na imagem original |
+| **Pré-visualização 3D ao vivo** | Gire, dê zoom e inspecione o avatar gerado em uma visualização Three.js antes de exportar |
+| **Exportação GLB em um clique** | Baixe um arquivo `.glb` padrão pronto para motores de jogos e visualizadores 3D |
+| **Funciona com ou sem servidor** | Análise completa via Gemini quando hospedado, fallback automático por canvas quando estático |
+
+## Arquitetura de modo duplo
+
+O GLB_FACTORY foi feito para rodar de duas formas diferentes dependendo de onde é implantado,
+e escolhe a correta automaticamente:
+
+1. **Modo com IA (hospedagem Node/Express)** — Em um ambiente completo, como
+ desenvolvimento local ou um contêiner na nuvem, o app conversa com um proxy de backend
+ ligado ao **Gemini 3.5 Flash**. O Gemini localiza automaticamente o rosto e extrai tom de
+ pele, cor do cabelo, cor da roupa e um penteado recomendado com alta precisão visual.
+
+2. **Modo estático de fallback (GitHub Pages)** — Sem backend disponível, o app detecta o
+ ambiente estático e muda para **análise facial no lado do cliente**: um leve amostrador
+ de canvas HTML5 lê os pixels do retrato diretamente no navegador e extrai as mesmas cores
+ de pele, cabelo e roupa, sem nenhuma requisição de rede.
+
+Mesma interface, mesmo resultado GLB, dois motores diferentes por trás, o que o ambiente de
+implantação conseguir realmente rodar.
+
+---
+
+## Início rápido
+
+```bash
+npm install
+npm run dev
+```
+
+Abra [http://localhost:3000](http://localhost:3000) no navegador.
+
+Para habilitar a análise com IA localmente, adicione uma chave do Gemini:
+
+```env
+# .env.local
+GEMINI_API_KEY=your_gemini_api_key_here
+```
+
+Sem a chave, o app continua funcionando, apenas usando o analisador de fallback no lado do
+cliente.
+
+## Implantação no GitHub Pages
+
+O repositório já vem com `.github/workflows/deploy.yml`, que constrói e publica o app
+estático a cada push para `main`.
+
+1. Acesse o repositório no GitHub, depois **Settings**.
+2. Em **Code and automation → Pages**, defina **Source** como **GitHub Actions**.
+3. Faça push para `main` e acompanhe o build na aba **Actions**.
+
+---
+
+## Tecnologias utilizadas
+
+| Camada | Stack |
+|---|---|
+| Renderização 3D | **Three.js** — renderização WebGL e construção procedural da malha do avatar |
+| Frontend | **React 19** + **Vite 6** — runtime e build da SPA |
+| Estilo | **Tailwind CSS v4** |
+| Ícones | **Lucide React** |
+| Backend | **Express** + **Google GenAI SDK** — proxy da API do Gemini |
+
+## Colaboradores
+
+- Angela — direção de produto, testes
+- Claude — implementação e fluxo de trabalho no GitHub
+
+## Aviso legal
+
+As fotos enviadas são processadas com o único propósito de gerar um avatar 3D. No modo com
+IA, os dados da imagem são enviados à API do Gemini sob os termos do Google; no modo estático
+de fallback, a análise acontece inteiramente no navegador e nada sai do dispositivo.
diff --git a/README.zh-CN.md b/README.zh-CN.md
new file mode 100644
index 0000000..3536a92
--- /dev/null
+++ b/README.zh-CN.md
@@ -0,0 +1,112 @@
+
+
+
