Skip to content

v0.5.0

Latest
Latest

Choose a tag to compare

View all tags
@hamzaamjad hamzaamjad released this 01 Jul 06:40
· 10 commits to master since this release
v0.5.0
44d1897

MCP Claude Context Server v0.5.0 Release Notes

Release Date: July 1, 2025

We're excited to announce the release of MCP Claude Context Server v0.5.0, a major upgrade that brings database storage, AI-powered search, and professional export capabilities to your Claude.ai conversation management.

🎯 Key Highlights

  • 10x Faster Search: SQLite FTS5 and FAISS-powered semantic search
  • Professional Exports: Obsidian, PDF, and bulk export capabilities
  • Easy Deployment: Docker support and one-click installers
  • Better Performance: 35% reduction in memory usage
  • Seamless Migration: Automated upgrade from v0.4.0

🚀 Major Features

1. Database Storage with SQLite

  • Migrated from JSON files to efficient SQLite database
  • Full-text search with FTS5 for instant results
  • Proper indexing for optimal query performance
  • ACID compliance for data integrity

2. AI-Powered Semantic Search

  • Find conversations by meaning, not just keywords
  • Powered by sentence-transformers and FAISS
  • Hybrid search combining text and semantic matching
  • Search suggestions and autocomplete

3. Professional Export Formats

  • Obsidian: Markdown with backlinks, tags, and daily notes
  • PDF: Professional documents with ToC and formatting
  • Bulk Export: Export multiple conversations at once
  • CSV/JSON: For data analysis and integration

4. Docker Deployment

  • Complete docker-compose setup
  • One command deployment: docker-compose up -d
  • Persistent volumes for data
  • Health checks and auto-restart

5. Enhanced MCP Tools

Six new tools for Claude Desktop:

  • semantic_search: AI-powered conversation search
  • export_to_obsidian: Create Obsidian vault exports
  • export_to_pdf: Generate PDF documents
  • bulk_export: Export multiple conversations
  • get_analytics: Retrieve usage statistics
  • find_related: Find related conversations

📊 Performance Improvements

Based on our testing with 100 conversations (2000 messages):

  • Database Operations: 1,433 conversations/second insertion
  • Text Search: 1.1ms average response time
  • Semantic Search: Sub-millisecond for indexed content
  • Memory Usage: Only 10MB overhead for search engine
  • Export Speed:
    • Obsidian: 1ms per conversation
    • PDF: 6ms per conversation

🔧 Technical Improvements

Compatibility

  • Python 3.11+ support (tested with 3.13)
  • SQLAlchemy 2.0 compatibility
  • Removed uvloop dependency for better compatibility
  • Fixed all deprecation warnings

Code Quality

  • Fixed SQL injection vulnerabilities
  • Proper error handling throughout
  • Comprehensive test suite
  • Performance benchmarking tools

Bug Fixes

  • Fixed metadata column naming conflict
  • Resolved row mapping issues in SQLAlchemy 2.0
  • Fixed SQL reserved keyword issues
  • Improved empty index handling

📦 Installation

Quick Start with Docker

cd deployment/docker
docker-compose up -d

Upgrade from v0.4.0

# Backup your data
cp -r extracted_messages extracted_messages_backup

# Run migration
poetry run python deployment/scripts/migrate_data.py

📚 Documentation

🧪 Testing

All components have been thoroughly tested:

  • ✅ Database operations and models
  • ✅ Search functionality (text and semantic)
  • ✅ Export formats (Obsidian, PDF)
  • ✅ Migration from v0.4.0
  • ✅ Docker deployment
  • ✅ Performance benchmarks

🔮 What's Next

Future enhancements planned:

  • Real-time conversation sync
  • Advanced analytics dashboard
  • More export formats (Notion, Roam)
  • Multi-language search support
  • Conversation clustering and insights

🙏 Acknowledgments

Thank you to all users who provided feedback on v0.4.0. Your input shaped these improvements.

📝 Breaking Changes

  • Database schema is new (automated migration provided)
  • Some internal APIs changed (MCP tools remain compatible)
  • Minimum Python version is now 3.11

🐛 Known Issues

  • Semantic search requires initial index building
  • PDF export is slower for very long conversations
  • Docker image is ~1GB due to ML models

📞 Support

Happy conversation managing! 🎉

The MCP Claude Context Team

Assets 2
Loading