🤖 Android AI Examples

A collection of 10 production-ready AI feature implementations for Android using Jetpack Compose, built step by step for developers who want to integrate AI into their Android apps.

Built with ❤️ by jinals-p

---

📱 Screenshots

Coming soon — screen recordings and screenshots will be added for each module.

---

🗂️ Modules

#	Module	Tech	Description	Status
01	QR / Barcode Scanner	ML Kit	Scan any QR code or barcode in real time	✅ Done
02	Face Detection	ML Kit	Detect faces, emotions, eye state and head direction	✅ Done
03	Text Recognition	ML Kit	Read and extract text from any real-world surface	✅ Done
04	Image Labeling	ML Kit	Identify and label objects in real time	✅ Done
05	Photo Explainer	Gemini API	AI describes any image in natural language	✅ Done
06	AI Chat	Gemini API	Full chatbot with streaming responses and memory	✅ Done
07	Voice Assistant	Gemini API	Talk to AI with your voice and hear it respond	⏳ Coming Soon
08	PDF Summarizer	Gemini API	Summarize any document with AI	⏳ Coming Soon
09	Custom AI Model	TFLite	Run your own trained AI model completely offline	⏳ Coming Soon
10	Complete AI App	All	Full production AI app combining everything	⏳ Coming Soon

---

⚡ Tech Stack

• Language: Kotlin
• UI: Jetpack Compose (100% — no XML)
• Architecture: MVVM + StateFlow + ViewModel
• AI/ML: ML Kit · Gemini API · TensorFlow Lite
• Camera: CameraX
• Navigation: Jetpack Navigation Compose
• Min SDK: 24 (Android 7.0)
• Target SDK: 35
• Theme: Custom dark blue/purple theme

---

🚀 Getting Started

1. Clone the repo

git clone https://github.com/jinals-p/AndroidAIExamples.git
cd AndroidAIExamples

2. Open in Android Studio

• Open Android Studio Hedgehog or later
• Select Open → choose the cloned folder
• Wait for Gradle sync to complete

3. Add your Gemini API key

Modules 05 and 06 require a free Gemini API key:

• Go to Google AI Studio
• Sign in with Google → Create API key → Copy it
• Find the API key constant in the project and replace:

const val GEMINI_API_KEY = "YOUR_API_KEY_HERE"

Free tier: 1,500 requests/day — no credit card needed

4. Run the app

• Connect your Android device or start an emulator
• Hit ▶️ Run in Android Studio
• The home screen lists all 10 modules — tap any card to open it
• Use the back button to return to the home screen

---

📦 Project Structure

AndroidAIExamples/
│
├── app/
│   └── MainActivity.kt                  → NavHost + HomeScreen + Screen routes
│
├── core/
│   ├── AppHeader.kt                     → Shared top bar component
│   ├── StatusBadge.kt                   → Shared badge component
│   └── theme/
│       ├── Color.kt                     → All color definitions
│       └── Theme.kt                     → Dark blue/purple theme
│
├── mlkitqrscanner/                      → Module 01
├── mlkitfacedetection/                  → Module 02
├── mlkittextrecognition/                → Module 03
├── mlkitimagelabeling/                  → Module 04
├── geminiphotoexplainer/                → Module 05
├── geminichatapp/                       → Module 06
├── geminivoice/                         → Module 07 (coming soon)
├── geminipdf/                           → Module 08 (coming soon)
├── tflitecustommodel/                   → Module 09 (coming soon)
└── completeaiapp/                       → Module 10 (coming soon)

---

🧭 Navigation

The app uses Jetpack Navigation Compose with a central NavHost in MainActivity.

// All available routes
Screen.Main           → "main_screen"            // Home screen
Screen.QrScanner      → "qr_scanner_screen"      // Module 01
Screen.FaceDetector   → "face_detector_screen"   // Module 02
Screen.TextDetector   → "text_detector_screen"   // Module 03
Screen.ImageLabeling  → "image_labeling_screen"  // Module 04
Screen.PhotoExplainer → "photo_explainer_screen" // Module 05
Screen.Chat           → "chat_screen"            // Module 06

Each screen receives an onBackCallback lambda to return home:

ScannerScreen(onBackCallback = {
    navController.navigate(Screen.Main.route)
})

---

🔍 Module Details

01 QR / Barcode Scanner

Package: com.app.aiexamples.mlkitqrscanner Tech: ML Kit Barcode Scanning + CameraX

Scan QR codes and barcodes in real time. Runs fully on-device — no internet required.

Features:

• Supports QR Code, EAN-13, EAN-8, Code 128, UPC-A, PDF417, Aztec and more
• Animated scan frame with color change on detection
• Torch / flashlight toggle
• Result card slides up on successful scan
• Scan again without restarting camera
• Camera permission with rationale and Settings redirect

Compose concepts: AndroidView · Canvas · AnimatedVisibility · DisposableEffect · rememberPermissionState

---

02 Face Detection

Package: com.app.aiexamples.mlkitfacedetection Tech: ML Kit Face Detection + CameraX

Detect faces in real time with bounding boxes, smile detection, eye state tracking, head direction and facial contour points.

Features:

• Live bounding box drawn exactly over detected faces
• Smile probability with animated progress bar
• Left/right eye open or closed detection
• Head direction — looking left, right or forward
• Facial contour overlay with 36+ landmark points
• Multiple faces simultaneously
• Front camera with correct left/right mirror compensation

Compose concepts: Canvas drawing · animateColorAsState · animateFloatAsState · KeepScreenOn · Front camera mirroring fix

---

03 Text Recognition

Package: com.app.aiexamples.mlkittextrecognition Tech: ML Kit Text Recognition + CameraX

Read and extract printed or handwritten text from any real-world surface in real time.

Features:

• Live text detection from camera feed
• Freeze / capture — stop scanning to read result clearly
• Copy to clipboard with animated success badge
• Share text via any installed app
• Text block count display
• Torch toggle for low-light conditions

Compose concepts: scaleIn/scaleOut animation · verticalScroll · LaunchedEffect with delay · Clipboard via Context

---

04 Image Labeling

Package: com.app.aiexamples.mlkitimagelabeling Tech: ML Kit Image Labeling + CameraX

Identify and label objects, scenes and concepts in real time using your camera.

Features:

• Real-time object identification with up to 8 labels shown
• Emoji mapping for all recognized labels
• Adjustable confidence threshold slider (30%–95%)
• Color-coded chips by confidence level
• Animated confidence progress bar with gradient
• Horizontal scrollable label chips

Compose concepts: LazyRow · Slider · animateFloatAsState · Horizontal gradient brush · Dynamic color theming

---

05 Photo Explainer

Package: com.app.aiexamples.geminiphotoexplainer Tech: Gemini API (gemini-1.5-flash) — multimodal vision

Take a photo or pick from gallery and ask Gemini AI anything about it.

Features:

• Take photo with camera or pick from gallery
• Custom prompt input
• Quick prompt suggestion chips
• Gemini Vision processes image + text together
• Loading state with spinner
• Sealed class state management (Idle / Loading / Success / Error)

Requires: Free Gemini API key from aistudio.google.com

Try these prompts:

"What is in this image?"
"Is this food healthy?"
"What brand is this product?"
"Translate any text visible in this image"
"Describe this for a visually impaired person"

Compose concepts: ActivityResultContracts · sealed class state · AnimatedVisibility · BitmapFactory

---

06 AI Chat

Package: com.app.aiexamples.geminichatapp Tech: Gemini API (gemini-1.5-flash) — streaming + chat history

A full AI chatbot with streaming responses word by word, conversation memory and a polished chat UI.

Features:

• Streaming responses — words appear one by one like ChatGPT
• Conversation memory across all messages in the session
• Animated typing indicator (3 bouncing dots)
• User and AI chat bubbles with different corner shapes
• Auto scroll to latest message
• Send via button or keyboard action
• Input disabled while AI is responding

Requires: Free Gemini API key from aistudio.google.com

Compose concepts: LazyColumn with key · StateFlow list updates · sendMessageStream · weight(1f) layout · Keyboard + ImeAction.Send handling · State hoisting

---

07 Voice Assistant

Package: com.app.aiexamples.geminivoice

⏳ Coming Soon

Speak to AI and hear it respond out loud. Uses Android built-in SpeechRecognizer and TextToSpeech with Gemini as the brain — no third-party libraries needed.

Planned features:

• Microphone pulse animation while listening
• Speech to text via Android SpeechRecognizer
• Gemini processes the spoken question
• AI response spoken aloud via TextToSpeech
• Full conversation history

---

08 PDF Summarizer

Package: com.app.aiexamples.geminipdf

⏳ Coming Soon

Pick any PDF document and get an AI-generated summary, key points and Q&A powered by Gemini.

Planned features:

• PDF file picker
• Page-by-page text extraction
• AI summary with key points
• Ask questions about the document

---

09 Custom AI Model

Package: com.app.aiexamples.tflitecustommodel

⏳ Coming Soon

Train a custom model in Python (Google Colab), export as .tflite and run it completely offline inside Android.

Planned features:

• Python training notebook included
• .tflite model bundled in assets
• Real-time inference on camera frames
• Runs 100% offline — no internet needed

---

10 Complete AI App

Package: com.app.aiexamples.completeaiapp

⏳ Coming Soon

A full production-ready AI app combining all previous modules with proper navigation, multi-screen MVVM architecture and a polished UI ready for Play Store.

---

🔑 API Keys

Module	Key needed	Where to get	Cost
01 QR Scanner	❌ None	—	Free
02 Face Detection	❌ None	—	Free
03 Text Recognition	❌ None	—	Free
04 Image Labeling	❌ None	—	Free
05 Photo Explainer	✅ Gemini	aistudio.google.com	Free
06 AI Chat	✅ Gemini	aistudio.google.com	Free
07 Voice Assistant	✅ Gemini	aistudio.google.com	Free
08 PDF Summarizer	✅ Gemini	aistudio.google.com	Free
09 Custom Model	❌ None	—	Free
10 Complete App	✅ Gemini	aistudio.google.com	Free

---

📚 What You Will Learn

Jetpack Compose

• LazyColumn / LazyRow with key for smooth animations
• AnimatedVisibility · animateColorAsState · animateFloatAsState
• Canvas drawing — bounding boxes, contours, scan overlays
• AndroidView — embedding CameraX PreviewView in Compose
• StateFlow + collectAsStateWithLifecycle
• Permission handling with Accompanist
• DisposableEffect · LaunchedEffect
• ViewModel + MVVM architecture
• NavHost + NavController navigation
• Keyboard and IME handling
• State hoisting pattern

Android AI

• ML Kit — on-device AI (no internet, no API key)
• CameraX — modern camera with ImageAnalysis
• Gemini API — multimodal AI (text + image + streaming)
• TensorFlow Lite — custom offline model inference
• Speech recognition + Text to speech
• Real-time frame-by-frame image analysis

---

🛠️ Requirements

• Android Studio Hedgehog (2023.1.1) or later
• Android device or emulator running API 24+
• Physical device recommended for camera features
• Internet connection required for Gemini modules (05–08)

---

🤝 Contributing

Contributions, issues and feature requests are welcome!

1. Fork the repo
2. Create your branch: git checkout -b feature/your-feature
3. Commit your changes: git commit -m 'Add some feature'
4. Push to the branch: git push origin feature/your-feature
5. Open a Pull Request

---

📄 License

MIT License — Copyright (c) 2025 jinals-p

Permission is hereby granted, free of charge, to any person obtaining
a copy of this software to use, copy, modify, merge, publish, distribute,
sublicense, and/or sell copies of the Software, subject to the condition
that the above copyright notice appears in all copies.

---

⭐ Support

If this repo helped you learn Android AI development, please give it a ⭐ star — it helps other developers discover this resource!

---

Made with ❤️ for the Android developer community