M.I.M.I.R - Multi-agent Intelligent Memory & Insight Repository

# Features **Explore NornicDB's advanced features and capabilities.** ## 🚀 Core Features ### Search & Discovery - **[Vector Search](../user-guides/vector-search.md)** - Semantic search with embeddings - **[Hybrid Search](../user-guides/hybrid-search.md)** - RRF fusion of vector + BM25 - **[Cross-Encoder Reranking](cross-encoder-reranking.md)** - Two-stage retrieval - **[Link Prediction](link-prediction.md)** - ML-based relationship prediction ### AI & Machine Learning - **[MCP Integration](mcp-integration.md)** - Model Context Protocol tools - **[Memory Decay](memory-decay.md)** - Time-based importance scoring - **[Vector Embeddings](vector-embeddings.md)** - Automatic embedding generation - **[GPU Acceleration](gpu-acceleration.md)** - 10-100x speedup - **[Auto-TLP](auto-tlp.md)** - Automatic relationship inference ### Configuration - **[Feature Flags](feature-flags.md)** - Runtime configuration - **[APOC Functions](apoc-functions.md)** - 450+ Neo4j-compatible functions ## 📚 Feature Categories ### Search Features NornicDB provides three complementary search methods: 1. **Vector Search** - Semantic similarity using embeddings 2. **Full-Text Search** - BM25 keyword matching 3. **Hybrid Search** - RRF fusion combining both [Learn more about search →](../user-guides/vector-search.md) ### AI Integration - Automatic embedding generation - MCP tool integration - Memory decay simulation - Link prediction - Auto-TLP (automatic relationship inference) [Learn more about AI features →](mcp-integration.md) ### Performance - GPU acceleration (Metal, CUDA, OpenCL) - Query caching - Index optimization - Parallel execution [Learn more about performance →](../performance/) ## 🎯 Popular Features ### Vector Search Search by meaning, not just keywords. Automatically generates embeddings for semantic search. ```cypher // Semantic search with string query CALL nornicdb.search.vector("machine learning algorithms", 10) YIELD node, score RETURN node.title, score ``` [Complete guide →](../user-guides/vector-search.md) ### GPU Acceleration 10-100x speedup for vector operations on Apple Silicon, NVIDIA, and AMD GPUs. ```yaml gpu: enabled: true backend: metal # or cuda, opencl ``` [Complete guide →](gpu-acceleration.md) ### Memory Decay Simulate human memory with time-based importance scoring. ```cypher // Get decay scores MATCH (n) RETURN n.title, n.decayScore ORDER BY n.decayScore DESC ``` [Complete guide →](memory-decay.md) ## 📖 Feature Guides - **[APOC Functions](apoc-functions.md)** - 450+ collection, text, math, graph functions - **[Auto-TLP](auto-tlp.md)** - Automatic relationship inference - **[Feature Flags](feature-flags.md)** - Runtime configuration - **[Link Prediction](link-prediction.md)** - Predict missing relationships - **[MCP Integration](mcp-integration.md)** - AI agent tools - **[Cross-Encoder Reranking](cross-encoder-reranking.md)** - Improve search accuracy ## ⏭️ Next Steps - **[User Guides](../user-guides/)** - Learn how to use features - **[API Reference](../api-reference/)** - Function documentation - **[Advanced Topics](../advanced/)** - Deep dive into internals --- **Explore features** → **[Vector Search](../user-guides/vector-search.md)**