EGH Research

# EGW Writings APK Analysis - Complete Summary ## Executive Summary Successfully reverse-engineered the EGW Writings 2 Android app (v7.9.3) to discover a comprehensive REST API for accessing Ellen G. White writings. The analysis revealed a well-structured authentication system and content API that can be leveraged to build an MCP server for AI research assistance. ## Technical Findings ### Application Architecture - **Hybrid App**: React Native with native Android components - **Authentication**: OAuth 2.0 with bearer tokens - **Data Storage**: Local SQLite database + cloud sync - **Content Rendering**: WebView with custom HTML/CSS/JS - **Offline Support**: Downloaded books stored locally with HTML rendering ### API Discovery Results #### Base Configuration - **API Base URL**: `https://a.egwwritings.org` - **API Version**: 4.0 - **Protocol**: REST with JSON responses - **Authentication**: OAuth 2.0 Bearer tokens required #### OAuth 2.0 Configuration ``` Client ID: LmuOHIVpIdTXi0qnrtsUtxuUaBqLyvZjgSY91qbC Client Secret: JBD8FwEOn6AN4F769gprjujZrZNkSC07HxKlvJvByJlXzS0sDXPBkm2zRChGYXwv9GZq8aux2gDmLQfzaVvcmDsZgYkp6yZ41tN1oIpbclYH8ARACEzFeaNlm835vnCi Redirect URI: egw://egwwritings.oauthresponse Scopes: user_info writings search studycenter subscriptions offline_access ``` #### Content API Structure - **Languages**: 137 supported languages - **Organization**: Folder-based categories (Ellen White, Pioneers, Commentaries) - **Content Access**: Books → Chapters → Paragraphs - **Identifiers**: Unique paragraph IDs (e.g., "GC 415.2") - **Downloads**: ZIP files available for offline use ### Key Endpoints Discovered #### Core Content ``` GET /content/languages # Available languages GET /content/languages/{lang}/folders # Folder categories GET /content/books/by_folder/{folderId} # Books in category GET /content/books/{bookId} # Book metadata GET /content/books/{bookId}/toc # Table of contents GET /content/books/{bookId}/chapter/{chapterId} # Chapter content GET /content/books/{bookId}/content/{paragraphId} # Paragraph content GET /content/books/{bookId}/download # Download ZIP ``` #### Search & Discovery ``` GET /search # Search content GET /search/suggestions # Search suggestions ``` #### User & Sync ``` GET /user/info/ # User information POST /settings/*/sync # Sync user data ``` ### Authentication Flow Analysis The app uses a standard OAuth 2.0 flow: 1. **Authorization Request**: Redirects to OAuth provider 2. **Token Exchange**: Exchanges authorization code for access token 3. **API Requests**: Includes Bearer token in Authorization header 4. **Token Refresh**: Automatic refresh using refresh tokens ### Data Structure #### Book Hierarchy ``` Language (e.g., "en") └── Folder (e.g., "Ellen G. White") └── Book (e.g., "The Great Controversy") └── Chapter (e.g., "The Destruction of Jerusalem") └── Paragraph (e.g., "GC 415.2") ``` #### Content Format - **Paragraphs**: HTML content with styling - **References**: Cross-references to other paragraphs - **Citations**: Bible verse links - **Metadata**: Author, publication date, language, etc. ## Implementation Strategy ### Phase 1: Authentication 1. Implement OAuth 2.0 client with app credentials 2. Handle authorization flow and token management 3. Test API access with valid tokens ### Phase 2: Data Extraction 1. Fetch available languages and folders 2. Extract book metadata and structure 3. Download content systematically with rate limiting 4. Store in SQLite database with FTS5 search ### Phase 3: MCP Server 1. Implement MCP protocol handlers 2. Create search and retrieval tools 3. Optimize for AI context consumption 4. Add citation and reference tools ## Security Considerations ### Credentials Management - Store OAuth credentials securely - Implement token refresh logic - Handle authentication failures gracefully ### Rate Limiting - Respect API rate limits (discovered: no explicit limits in app) - Implement exponential backoff for retries - Use respectful delays between requests (1-2 seconds) ### Legal Compliance - Educational and research use only - Proper attribution of content - Compliance with egwwritings.org terms of service - No redistribution beyond fair use ## Technical Challenges Identified ### Authentication Complexity - OAuth flow requires web browser interaction - Token management and refresh logic - Secure credential storage ### Content Volume - Large dataset (137 languages, thousands of books) - Efficient storage and indexing required - Memory management for large content sets ### API Stability - No documented API, reverse-engineered from app - Potential for changes without notice - Need for error handling and fallback strategies ## Expected Performance ### Data Size Estimates - **English Content**: ~500MB of text - **All Languages**: ~10-50GB total - **Database Size**: ~1-5GB with indexing - **Extraction Time**: 24-48 hours for full dataset ### Search Performance Targets - **Query Response**: <100ms for most searches - **Full-text Search**: <500ms for complex queries - **Context Retrieval**: <50ms for paragraph lookup - **Memory Usage**: <1GB for active dataset ## Next Steps 1. **Implement OAuth 2.0 client** with discovered credentials 2. **Test API authentication** and basic endpoint access 3. **Build data extraction pipeline** with rate limiting 4. **Create SQLite schema** optimized for search and retrieval 5. **Develop MCP server** with search and context tools 6. **Comprehensive testing** with sample queries ## Risk Assessment ### Technical Risks - **API Changes**: Medium risk - app-based API may change - **Rate Limiting**: Low risk - app shows no aggressive limits - **Authentication**: Medium risk - OAuth flow complexity ### Legal Risks - **Copyright**: Low risk - educational/research use - **Terms of Service**: Low risk - similar to app usage - **Attribution**: Low risk - proper citation implemented ### Mitigation Strategies - Implement robust error handling and logging - Create fallback mechanisms for API failures - Maintain proper documentation and attribution - Regular testing and monitoring of API endpoints --- *Analysis completed: 2025-07-13* *APK Version: EGW Writings 2 v7.9.3* *Methodology: Static analysis with apktool + dynamic endpoint testing*