This is an Obsidian MCP that utilizes RAG, GPU acceleration, HNSW, and multi-query functionality. This tool is designed for heavy users who store and utilize numerous notes in Obsidian. It enables instantaneous semantic search (superior to keyword search) of Obsidian notes and PDFs from Claude.
Even Python beginners can easily install it, and detailed installation instructions are provided in the readme.
Installation link: (Link insert is not allowed… why?)
Features
Core Search Capabilities
- Hybrid Search Engine: Advanced vector similarity + keyword search fusion
- Intelligent Semantic Search: High-precision meaning-based document retrieval
- Adaptive Query Processing: Automatic parameter adjustment based on query complexity
- Multi-modal Search: Integrated text and attachment file search
- Contextual Expansion: Related document discovery and context-aware retrieval
Advanced AI & RAG Features
- Hierarchical Retrieval: Document → Section → Chunk progressive search
- Multi-query Fusion: Intelligent combination of multiple search queries with weighted averaging, maximum value, and reciprocal rank fusion
- Adaptive Chunk Retrieval: Dynamic chunk size adjustment based on query complexity
- Knowledge Graph Exploration: Vector similarity-based connection discovery with BFS traversal and graph centrality ranking
- Temporal-aware Search: Balance between relevance and recency with time-weighted scoring
Advanced Metadata Filtering
- Complex Tag Logic: AND/OR/NOT combinations for sophisticated tag-based filtering
- Time Range Filtering: Precise temporal document filtering
- File Type & Quality Filtering: Content quality assessment and file type categorization
- Multi-dimensional Filtering: Simultaneous application of multiple filter criteria
Performance Optimization
- GPU/CPU Auto-selection: Hardware-optimized index selection (GPU_IVF_FLAT, GPU_CAGRA, HNSW)
- HNSW Index Optimization: Dynamic parameter tuning (ef, nprobe) based on collection size
- Real-time Performance Monitoring: Live performance tracking and analysis
- Adaptive Search Parameters: Result quality-based parameter dynamic adjustment
- Batch Processing Optimization: Efficient large-scale search operations
System Intelligence
- Auto-tuning: Collection size-based automatic parameter optimization
- Performance Benchmarking: Multi-strategy search performance comparison
- Smart Recommendations: Automatic optimization suggestions based on usage patterns
- Resource Management: Intelligent memory and GPU utilization
Integration & Connectivity
- Claude Desktop Integration: Seamless connection via FastMCP protocol
- Real-time File Monitoring: Automatic change detection and incremental reindexing
- Multilingual Support: Advanced embedding models for global language support
- Container-based Deployment: Reliable Podman-based Milvus deployment