You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Obsidian MCP-RAG server for AI agents [Local & Cloud] with full-note, heading-aware semantic retrieval, lexical search, graph navigation, passage provenance, and vault manipulation tools with a live-monitoring web-dashboard.
This is the part 2 of a GraphRAG system, in which the user interacts with the data through 2 data structures: vector database (Pinecone) and graph database (Neo4j). I leverage the probabilism of vector embedding and the determinism of a knowledge graph to minimize hallucination and maximise explainability. The domain: ‘Electronic’ music genre.
This is the part 1 of a GraphRAG system, which implements the Data Pipeline orchestrated by Dagster, where the data is ingested from multiple sources for building a knowledge graph in Neo4j Aura and an embedding model in Pinecone. The domain: ‘Electronic’ music genre. Contact me to see a demo.