Zerch

Logs are hard to search because traditional tools rely on exact keywords and miss semantically related issues. Different wording for the same problem makes debugging slow and inefficient. Existing vector databases are often heavy for lightweight log workflows. This project solves the problem by introducing a fast, lightweight semantic search engine for logs. Built as a Rust package, it can be used as a CLI tool or integrated into AI agents and CI pipelines. It converts logs into embeddings and retrieves results using similarity instead of keywords.