English | Русский

Storog is an Android application that transforms your device into an intelligent visual monitoring system. It uses the device's camera to detect motion in a designated area. Upon detecting motion, Storog utilizes the Gemini AI model to analyze the scene and can send notifications, including an image and AI analysis, to a specified Telegram chat.

• Visual Change Detection: Monitors the camera feed for changes compared to a baseline image.
• Customizable Sensitivity: Allows users to set a threshold for how significant a change needs to be to trigger an alert.
• AI-Powered Image Analysis: Leverages Google's Gemini model to analyze captured images based on a user-defined prompt (e.g., "Is there a person in the image?", "Is the door open?").
• Telegram Notifications: Sends alerts to a specified Telegram chat, including the captured image and the AI-generated description or analysis.
• Customizable AI Prompts: Users can tailor the text prompt sent to the AI to focus the analysis on specific aspects of the image.
• Settings Interface: Provides an in-app settings screen for configuring the sensitivity threshold, AI prompt, and Telegram chat ID.
• Camera Preview: Displays the live camera feed within the app.
• Start/Stop Control: Simple controls to activate or deactivate the monitoring process.

Storog is an Android application built using Gradle.

• Android Studio (latest stable version recommended).
• An Android device or emulator with camera capabilities.

To use the AI analysis and Telegram notification features, you need to configure the following:

1. Google Gemini API Key:
   • Obtain an API key from Google AI Studio.
2. Telegram Bot Token:
   • Create a new Telegram bot by talking to BotFather on Telegram.
   • BotFather will provide you with an API token.
3. Telegram Chat ID:
   • This is the ID of the chat, group, or channel where the bot should send notifications.
   • You can get your personal chat ID by starting userinfobot on Telegram. For groups or channels, other methods might be needed (e.g., temporarily adding a bot that reveals chat IDs).

Configuration File:

• After obtaining your Gemini API key and Telegram Bot token, create a file named my_config.properties in the app/src/main/assets/ directory of your project.

• Add your keys to this file in the following format:

  GEMINI_API_KEY=YOUR_GEMINI_API_KEY_HERE
  MY_BOT_TOKEN=YOUR_TELEGRAM_BOT_TOKEN_HERE

  Note: Replace YOUR_GEMINI_API_KEY_HERE and YOUR_TELEGRAM_BOT_TOKEN_HERE with your actual keys.

• Important: Add my_config.properties to your .gitignore file to avoid accidentally committing your API keys to version control.

  # API keys
  app/src/main/assets/my_config.properties
  

Setting the Target Chat ID:

• The Telegram Chat ID is configured within the app itself.
• Navigate to the "Settings" screen in the app to enter and save your target Chat ID.

1. Install the App: Build and run the application on your Android device or emulator.
2. Grant Permissions: On the first launch, the app will request camera permission. You must grant this permission for the app to function.
3. Configure Settings:
   • Navigate to the "Settings" screen.
   • Enter your Telegram Chat ID where you want to receive notifications.
   • Optionally, you can adjust the Change Threshold (0-100%). A lower value means higher sensitivity to changes. Default is 5%.
   • Optionally, you can customize the AI Prompt to guide the image analysis. Default is "is there a cat in the picture?".
   • Save your settings.
4. Start Monitoring:
   • Return to the main screen.
   • Tap the "Start" button.
   • The app will capture an initial reference image. Monitoring is now active.
5. Monitoring in Action:
   • The app will continuously compare the current camera view with the reference image.
   • The current difference percentage is displayed.
   • If the difference exceeds your set threshold, the app will:
     • Send the current image and your AI prompt to Gemini AI for analysis.
     • Send a message to your configured Telegram chat, including the image and the AI's response (unless the AI response dictates otherwise, e.g., if the response starts with "No", or if a message limit is reached).
6. Stop Monitoring: Tap the "Stop" button to deactivate monitoring.
7. Help: Tap the "Help" button for a brief in-app guide.

• More Secure API Key Management: Instead of a properties file in assets, consider using the Android Keystore system or passing keys via Gradle build parameters from a local.properties file (which should also be in .gitignore).
• Background Operation: Currently, monitoring is only active when the app is in the foreground. A background service could enable continuous monitoring.
• Advanced Scheduling: Allow users to schedule monitoring for specific times or days.
• Notification Customization: More options for notification sounds, vibration, etc.
• Multi-Camera/Scene Support: Ability to configure and switch between different monitored areas or camera views if the device has multiple cameras.
• Error Handling and Resilience: More robust handling of network errors, API failures, etc.
• Internationalization: Translate UI elements into more languages (currently, there's a mix of English and Russian in the source code/UI descriptions).
• Message Limit: The app currently has a hardcoded limit of 3 Telegram messages before stopping monitoring to prevent excessive notifications. This could be made configurable or use a more sophisticated rate-limiting approach.
• AI Response Parsing: The logic for skipping Telegram messages (e.g., if the AI says "No") or stopping monitoring is based on simple string prefixes. This could be made more flexible.

1. Clone the Repository:

   git clone https://github.com/premananda108/storog.git
   cd <repository-directory>

2. Configure API Keys: Ensure you have set up the my_config.properties file in app/src/main/assets/ as described in the "Installation and Configuration" section.
3. Open in Android Studio: Open the root project directory in Android Studio.
4. Sync Gradle: Allow Android Studio to sync Gradle files and download dependencies.
5. Build the Project:
   • You can build the project using the "Build" menu in Android Studio (e.g., "Build" > "Make Project" or "Build" > "Build Bundle(s) / APK(s)").
   • Alternatively, you can build from the command line using the Gradle Wrapper:
     • For a debug APK: ./gradlew assembleDebug
     • For a release APK (requires release signing configuration): ./gradlew assembleRelease
6. Run: Deploy the app to an Android device or emulator.

This project is licensed under the MIT License - see the LICENSE file for details.