Can I run this locally for privacy?

Yes, this skill provides detailed configurations for using Ollama with models like nomic-embed-text, ensuring your data remains 100% local and private.

How does it handle long markdown files?

The skill includes markdown-aware and semantic chunking strategies that split long documents into smaller, meaningful pieces, preserving context while staying within model token limits.

What is the advantage of vector search over keyword search?

Vector search understands the meaning and context of your notes, allowing you to find relevant content even when exact keywords don't match, such as finding 'habit formation' when searching for 'routine building'.

Does it support hybrid search?

Yes, it provides implementation patterns for combining traditional keyword search (BM25) with semantic vector scores to provide the most accurate search results.

What embedding models are supported?

The skill supports local models via Ollama (nomic-embed-text, all-minilm) as well as cloud-based APIs from OpenAI and Voyage AI.

Vector Search for Zettelkasten

Name: Vector Search for Zettelkasten
Author: leobrival

byleobrival

Ciencia de Datos y ML

Implements semantic search and AI-driven discovery for Obsidian vaults using vector embeddings and advanced chunking strategies.

Acerca de

Vector Search for Zettelkasten transforms your Obsidian vault into a semantically searchable knowledge base. By converting markdown notes into vector embeddings, it moves beyond keyword matching to understand the conceptual relationships between ideas. This skill provides a comprehensive framework for implementing local or API-based embedding models, optimized chunking for long documents, and hybrid search capabilities. It is particularly useful for PKM (Personal Knowledge Management) enthusiasts looking to automate link suggestions, generate Maps of Content (MOCs), and discover hidden connections across thousands of notes without manual tagging.

Características Principales

Hybrid search combining keyword matching with cosine similarity scoring
Automated link suggestions and Map of Content (MOC) structure generation
Advanced markdown-aware and semantic chunking strategies for precise retrieval
0 GitHub stars
Support for local (Ollama) and API-based (OpenAI, Voyage AI) embedding models
Incremental indexing with efficient change detection and batch processing

Casos de Uso

Finding conceptually related notes in a large Zettelkasten without manual tagging
Building a private, local-first RAG (Retrieval-Augmented Generation) system for personal notes
Generating structured Maps of Content for complex research topics automatically

Acerca de

Características Principales

Hybrid search combining keyword matching with cosine similarity scoring
Automated link suggestions and Map of Content (MOC) structure generation
Advanced markdown-aware and semantic chunking strategies for precise retrieval
0 GitHub stars
Support for local (Ollama) and API-based (OpenAI, Voyage AI) embedding models
Incremental indexing with efficient change detection and batch processing

Casos de Uso

Finding conceptually related notes in a large Zettelkasten without manual tagging
Building a private, local-first RAG (Retrieval-Augmented Generation) system for personal notes
Generating structured Maps of Content for complex research topics automatically