ReaME TTS Storage Architecture

Overview

AudioFlow Reader implements a local-first architecture for text-to-speech storage, optimizing for reliable offline playback while preparing for future cloud synchronization.

Storage Strategy

Local Storage (IndexedDB)

• Database: TextToSpeechDB (Dexie.js implementation)

• Current Version: 3

• Primary Storage: Audio blobs stored directly in IndexedDB

• Size Limits: 50MB per audio file (configurable)

Data Schema

Copyinterface TextToSpeechEntry {
  id?: number;          // Auto-incrementing primary key
  text: string;         // Original text content
  audioBlob: Blob;      // Compressed audio data
  size: number;         // Storage size tracking
  timestamp: Date;      // Creation timestamp
  sourceType: 'url' | 'paste';  // Content source
  sourceUrl?: string;   // Original URL if applicable
  imageUrl?: string;    // Associated image URL
  syncStatus: 'pending' | 'synced';  // Cloud sync status
}

Indexing Strategy

• Primary Index: Auto-incrementing ID

• Secondary Indexes:

  • timestamp: For chronological queries

  • sourceType: For filtering by content source

  • syncStatus: For cloud sync management

Design Decisions

1. Blob Storage vs Base64

• Decision: Store audio as Blob instead of Base64

• Rationale:

  • ~25% smaller storage footprint

  • Faster retrieval for playback

  • No encoding/decoding overhead

  • Native browser handling

2. Local-First Architecture

• Decision: Prioritize local storage with sync preparation

• Rationale:

  • Reliable offline playback

  • Better user experience

  • Reduced bandwidth usage

  • Future-proof for cloud sync

3. Size Management

• Decision: Implement 50MB per-file limit

• Rationale:

  • Prevents IndexedDB quota issues

  • Reasonable size for speech content

  • Balances quality and storage

Architecture Decisions

1. Connection Management

• Decision: Use per-operation database connections instead of a global instance

• Rationale:

  • Prevents issues with Promise cloning in IndexedDB

  • Ensures proper cleanup of database resources

  • Avoids state management complexity

• Implementation: Each database operation creates its own connection and closes it after use

2. Audio Data Storage

• Decision: Store audio as ArrayBuffer instead of Blob

• Rationale:

  • ArrayBuffers are more efficiently stored in IndexedDB

  • Avoids DataCloneError issues with Blob objects

  • Maintains compatibility with Web Audio API

• Implementation: Convert audio data to ArrayBuffer before storage, create Blob only when needed for playback

3. Storage Monitoring

• Decision: Use polling instead of database hooks for storage monitoring

• Rationale:

  • More reliable than hooks with per-operation connections

  • Simpler implementation

  • Regular updates without complex event handling

• Implementation: Poll storage usage every 5 seconds in the StorageIndicator component

4. Error Handling

• Decision: Implement comprehensive error handling at all database operations

• Rationale:

  • Ensures database connections are always properly closed

  • Provides clear error messages for debugging

  • Maintains application stability

• Implementation: Try-catch blocks with connection cleanup in finally blocks

Future Enhancements

1. Audio Compression

• Implement Web Audio API compression

• Add configurable quality settings

• Optimize storage-quality trade-off

2. Cloud Sync (Supabase)

• Implement background sync

• Add conflict resolution

• Manage bandwidth usage

• Track sync status

3. Storage Management

• Implement automatic cleanup

• Add user storage preferences

• Monitor and manage quotas

Technical Considerations

Browser Compatibility

• IndexedDB support: All modern browsers

• Storage limits: Browser-dependent

• Blob handling: Standardized across platforms

Performance

• Async operations for large files

• Compressed storage for efficiency

• Direct blob playback for performance

Security

• Local-only storage currently

• Prepared for secure cloud sync

• No sensitive data exposure

Migration Strategy

• Version control through Dexie

• Automatic schema updates

• Data preservation during upgrades