Show HN: Jev Powered Obsidian Search (github.com)

Jev Graph Search has introduced a new AI tool designed to enhance the search capabilities within local Obsidian vaults and Logseq Markdown notes. This tool allows users to find relevant notes and passages efficiently, generating a shortlist on their machine and then employing Jev's ranking service to ensure high-quality results. Significantly, the tool has shown impressive performance improvements, such as raising the recall rate of top-five sources from 63.9% to 81.8% on an extensive 8,851-node tax code graph. In addition, all results retain original source references, creating a seamless link between queries and the underlying data. For technical users, Jev Graph Search operates through a command-line interface (CLI) and requires Node.js 20+ and npm for installation. Upon setup, users can search Markdown directories interactively or execute persistent commands, enabling direct retrieval of information, proposal of new notes, and auditing of existing structures. The tool also maintains robust security practices by managing API keys securely and offering offline functionality for lexical ranking. This development could greatly impact how researchers and professionals interact with their note-taking systems, transforming fragmented data into a cohesive and searchable resource.