Release Notes - Version 0.1.3-alpha

CRITICAL BREAKING CHANGES

This release includes a complete database migration from SQLite/ChromaDB to PostgreSQL with pgvector.

Required Actions for Upgrading:

1. Backup any important documents - All data will be lost during migration
2. Run docker compose down -v to remove old volumes
3. Re-run the setup script to generate new .env with secure passwords
4. Start fresh with docker compose up -d

Major Features

PostgreSQL Database Migration

• Complete migration from SQLite to PostgreSQL with pgvector extension
• Improved concurrency handling and performance

Enhanced Document Processing

• Fixed concurrent document ingestion issues
• Improved CLI document management with proper cleanup commands
• Document deduplication using SHA256 hashing

User Settings Integration

• All research parameters now properly respect user settings
• Dynamic configuration for:
  • Planning context limits (critical for local LLMs)
  • Note content windows
  • Writing preview sizes
  • Search result counts
  • Concurrent operation limits
• Per-mission configuration support

New Features

CLI Improvements

• New cleanup command to remove failed documents: maestro-cli.sh cleanup
• Enhanced parallel document processing with configurable batch sizes
• Better error handling and recovery
• Progress tracking for bulk operations

CPU Mode Enhancements

• Dedicated docker-compose.cpu.yml for CPU-only deployments
• Automatic hardware detection and optimization
• Optimized batch sizes based on available hardware

Writing Mode Enhancements

• Deep search functionality for comprehensive research
• Improved reference management
• Better draft synchronization
• Enhanced WebSocket stability

Research Mode Improvements

• Fixed planning agent context overflow issues

Bug Fixes

Critical Fixes

• Fixed user settings not being respected (using hardcoded values instead of dynamic config)
• Resolved concurrent chat blocking issues
• Fixed CLI document creation timing problems
• Corrected WebSocket connection stability issues
• Fixed writing mode state management

UI/UX Fixes

• Improved document upload progress tracking
• Better error messages and user feedback
• Fixed settings modal saving issues
• Corrected document filter persistence

Performance Improvements

• Reduced database connection overhead with NullPool configuration
• Optimized vector search with hybrid dense/sparse retrieval
• Improved document chunking and embedding efficiency
• Better memory management for large documents
• Reduced WebSocket message overhead

Configuration Changes

Security Improvements

• Secure password generation in setup scripts
• JWT secret configuration
• Improved admin password reset functionality
• Better credential management via environment variables

Platform-Specific Notes

Windows

• Automatic line ending fixes
• Better Docker Desktop integration

Linux

• Optimized for both CPU and GPU deployments

Known Issues

• Large PDF processing may require reduced batch sizes on systems with limited memory
• Local LLMs with small context windows require manual parameter adjustment
• Some AMD GPUs may need CPU mode due to ROCm limitations

Upgrade Instructions

1. Stop current services:

   docker compose down

2. Remove old volumes (WARNING: Data loss):

   docker compose down -v

3. Update repository:

   git pull

4. Generate new configuration:

   ./setup-env.sh  # or setup-env.ps1 on Windows

5. Start services:

   ./start.sh # or docker-compose.cpu.yml for CPU mode

6. Wait for initialization:

   If you're having trouble logging in, wait for the following message in the logs.

   docker compose logs -f maestro-backend
   # Wait for "Application startup complete"

Recommendations

For Local LLM Users

• Adjust Planning Context in Settings → Research Parameters
• Set to 30,000-40,000 for 8K context models
• Set to 100,000-120,000 for 32K context models

For CPU-Only Users

• Use docker-compose.cpu.yml for optimal performance
• Reduce batch sizes in environment configuration
• Consider limiting concurrent document processing

For detailed upgrade instructions and troubleshooting, please refer to:

• README.md
• TROUBLESHOOTING.md
• WINDOWS_SETUP.md