SmartMediaAI – Video Summarizer App using Firebase Gemini + Jetpack Compose

SmartMediaAI is an Android app that lets users select a video, summarize it using Gemini via Firebase, and listen to the summary using Text-to-Speech. Works with YouTube links and direct video URLs.

Features

• In-App Video Player: Built-in player using Jetpack Media3 (ExoPlayer) to handle YouTube and direct video links.
• AI-Powered Summaries: Generates smart summaries of video content using the Firebase Gemini API.
• Text-to-Speech Playback: Listen to the generated summaries with controls for different accents.
• Modern UI: A clean, responsive interface built entirely with Jetpack Compose.
• Predefined Video List: A dropdown menu to easily select from a list of sample videos.

Project Structure

Here is an overview of the key files and directories in the project:

com.anandgaur.smartmediaai  
├── MainActivity.kt                          # Entry point of the app  
├── player/                                  # Video playback components  
│   ├── VideoPlayer.kt                       # Jetpack Compose-based video player  
│   └── VideoSelectionDropdown.kt            # UI for selecting videos  
├── screen/  
│   └── VideoSummarizationScreen.kt          # Main UI screen for summarization  
├── ui/                                      # UI components for output  
│   ├── TextToSpeechControls.kt              # Controls for TTS playback  
│   └── OutputTextDisplay.kt                 # Displays the summarized output  
├── util/                                    # Utility classes and helpers  
│   └── VideoItem.kt                         # Data model for video entries  
├── viewmodel/                               # ViewModel and state management  
│   ├── VideoSummarizationViewModel.kt       # Handles UI logic and API interactions  
│   └── OutputTextState.kt                   # UI state model for summarized output  
└── SmartMediaAIApplication.kt               # Application class for global setup

Code Overview

This project is follows the MVVM (Model-View-ViewModel) architecture for a clean separation of concerns and Jetpack libraries for best practices. Here's a breakdown of how the app works, file by file.

1. The Starting Points

• SmartMediaAIApplication.kt

  • What it is: The very first code that runs when the app is launched, even before any screen appears.
  • What it does: Performs one-time setups for the entire application, initializing Firebase and setting up Hilt for dependency management.

• MainActivity.kt

  • What it is: The main "window" or entry point for the user interface.
  • What it does: Its only responsibility is to load the overall theme and display our main screen, VideoSummarizationScreen.

2. The Main Screen: VideoSummarizationScreen.kt

• What it is: The heart of our user interface, where the user interacts with the app.
• What it does: It acts as a manager for all visual components.
  • Assembles the UI: Brings together the video player, dropdown menu, "Summarize" button, and the text area for the output.
  • Manages State: Keeps track of important information, like which video is currently selected.
  • Handles Video Players: Decides which video player to show. It uses a special YouTube player for YouTube links and our custom VideoPlayer for other links (like .mp4 files).
  • Connects to the "Brain": Tells the ViewModel to start working when the user clicks the "Summarize" button.
  • Manages Text-to-Speech: Initializes and controls the Android Text-to-Speech engine.

3. The "Brain" of the App: The ViewModel

• VideoSummarizationViewModel.kt

  • What it is: The "brain" of our app. It handles all logic and heavy lifting behind the scenes, separate from the UI.
  • What it does: Takes the video link, creates a request for the Firebase Gemini AI, and sends it to Google's servers. As it gets a response, it updates its status, which the UI automatically reflects.

• OutputTextState.kt

  • What it is: A simple file that defines the possible states of our summarization process (Initial, Loading, Success, Error).
  • What it does: It tells the UI what's happening, allowing it to intelligently show a loading spinner, an error message, or the final summary.

4. The Reusable UI Components

• VideoPlayer.kt (in the player package)

  • What it is: Our custom video player for standard video files (not YouTube).
  • What it does: Uses Google's powerful ExoPlayer to play videos and automatically shows a loading spinner when buffering.

• VideoSelectionDropdown.kt (in the player package)

  • What it is: The dropdown menu that lets the user pick a video.
  • What it does: Displays a list of our sample videos and informs the main screen of the user's choice.

• OutputTextDisplay.kt (in the ui package)

  • What it is: The text box at the bottom that shows the summary.
  • What it does: Displays text based on the OutputTextState from the ViewModel. The success text is scrollable.

• TextToSpeechControls.kt (in the ui package)

  • What it is: The UI for the "Listen" feature.
  • What it does: Provides the "Listen" / "Pause" button and the accent selection dropdown.

5. The Helpers and Data: The util package

• VideoItem.kt & SampleVideoList.kt

  • What they are: A data blueprint for a video and a predefined list of sample videos for the dropdown menu.

• extractYouTubeVideoId.kt

  • What it is: A specialized helper function.
  • What it does: Its only job is to take a full YouTube URL and pull out the unique 11-character video ID needed by the player.

Tech Stack

Architecture	MVVM Architecture
Language	Kotlin
UI	Jetpack Compose
Dependency Injection	Hilt
Jetpack Media3	video playback
Firebase	Gemini AI APIs

Demo

Follow and Connect with Me:

• 𝗚𝗶𝘁𝗛𝘂𝗯
• 𝗟𝗶𝗻𝗸𝗲𝗱𝗜𝗻
• 𝗬𝗼𝘂𝗧𝘂𝗯𝗲
• 𝗜𝗻𝘀𝘁𝗮𝗴𝗿𝗮𝗺

Setup Instructions

1. Clone the repository:

   git clone https://github.com/anandgaur22/SmartMediaAI.git
   

2. Add your google-services.json in the app/ directory.

3. Enable Gemini API and Vertex AI Gemini API from Firebase > Build > Generative AI.

4. Sync Gradle and run the project on a physical/emulator Android device.

Read More

• AI Use Cases in Jetpack Media3 Playback in Android
• Advance Android Masterclass: Real Examples

❤️ Support This Project

If you've found value in the Android Development in-depth industry-ready course, consider supporting the effort and dedication poured into its creation. Your contribution on Buy Me a Coffee will fuel the creation of more content, enable continuous improvement, and help build a community of motivated learners. It's not just about a cup of coffee; it's about fostering a culture of support and collaboration. Together, we can create something exceptional. Your support is not just appreciated; it's a cornerstone for the future of quality education. Thank you for being a part of this incredible journey!

Badges