A Flutter Android app that automatically organizes your lecture photos by subject, lets you search every word inside them, and generates AI study guides instantly. Screenshot a slide, a ChatGPT explanation, a textbook page, or a whiteboard, LectureVault reads it and files it into the right folder instantly.

You screenshot lecture slides, type out notes, grab explanations from ChatGPT, photograph textbook pages and all of it ends up in one chaotic camera roll. Finding anything later is a nightmare.

LectureVault fixes that. Every photo goes into the right subject folder automatically, without you having to think about it.

Digital screenshots are where LectureVault shines:

• Lecture slides and presentation screenshots
• ChatGPT / AI tool explanations
• Textbook and PDF page photos
• Typed notes and study guides
• Website and article screenshots

Printed and neat handwritten notes are also well supported. The OCR pipeline handles clean text reliably.

Heavily cursive or very messy handwriting is best effort. Classification still works when enough keywords are readable, but digital content gives the most consistent results.

Text recognition runs entirely on your device. No internet required for this step.

It extracts every single word from your images using background OCR. An enhancer layer then:

• fixes common OCR mistakes
• removes noise words
• extracts meaningful keywords
• detects academic patterns using regex

All OCR results from a batch of photos are sent to Gemini in a single request.

Gemini returns:

• subject label
• confidence score

The prompt is structured so Gemini reasons through the content before assigning a label, improving accuracy on tricky or mixed content.

Every word from your screenshots is saved into a local search index under a dedicated ocrFull field.

• Supports multi-word queries
• Matches words anywhere in the image
• Returns results instantly

Deleting a photo also removes it from the index to keep everything clean and lightweight.

You can select:

• individual images
• or full subject folders

The app sends extracted OCR text to Gemini and generates:

• summaries
• flashcards

No prompt writing needed. It just works.

• Photos are stored in subject-based folders
• Folder location is user-defined
• Preferences stored using SharedPreferences

Nothing leaves your device except OCR text sent to Gemini for classification.

You can:

• browse notes
• zoom into images
• search instantly
• share directly

All inside the app.

Includes:

• animated splash screen
• search tab in bottom navbar
• improved UI with vivid buttons

Think of it as your own notes app that organizes everything for you automatically.

• Flutter (Android)
• Google ML Kit Text Recognition
• Gemini Flash API (free tier)
• SharedPreferences
• image_picker
• file_picker
• permission_handler
• path

• Flutter SDK
• Android physical device (API 21+)
• Gemini API key (free at https://ai.google.dev)

git clone https://github.com/Aneezakiran07/LectureVault

cd lecturevault
flutter pub gett

Create a .env file in the project root:

GEMINI_API_KEY=your_key_here

Add .env to pubspec.yaml:

flutter:
   assets:
    -  .env

Run on a connected device:

flutter run

lib/
├── main.dart
├── core/constants/
│   └── colors.dart
├── models/
│   ├── photo_item.dart
│   └── subject_folder.dart
├── screens/
│   ├── splash_screen.dart
│   ├── onboarding/folder_setup_screen.dart
│   ├── home/home_screen.dart
│   ├── upload/upload_screen.dart
│   ├── folder_view/folder_view_screen.dart
│   ├── search/search_screen.dart
│   └── settings/settings_screen.dart
└── services/
    ├── storage_service.dart
    ├── permission_service.dart
    ├── search_service.dart
    ├── ocr_service.dart
    ├── ocr_enhancer.dart
    ├── ocr_result.dart
    ├── classifier_service.dart
    └── gemini_classifier.dart
gemini_classifier.dart

LectureVault handles async operations and edge cases carefully:

• Memory Management Background processes stop instantly when leaving screens using mount guardss

• Batch Upload Safety Uses hash-based filenames to prevent overwrite during fast processingg

• Smart Error Handling Duplicate and constraint errors are treated as success to avoid crashess

• Index Sanitization Filters out UI error messages so search data stays cleann

Gemini free tier allows 500 requests per day.

LectureVault batches uploads:

• 1 session = 1 request
• 20 photos = still 1 request

That is up to ~10,000 photos per day.

• Android only
• Low-text images may be marked Unclassified
• Internet required for Gemini classification
• OCR works offline

Built incrementally:

UI → Storage → File handling → Folder view → Settings → OCR → Classification → Search → AI study generation → Async + memory fixes

Each stage was developed and improved step by step.