Zvec
A lightweight, in-process vector database that embeds directly into an application for similarity search, combining dense and sparse vectors with full-text search and structured filters.
Zvec is an in-process vector database built for embedding directly into applications rather than running as a separate server. It supports dense and sparse embeddings, multi-vector queries, and a range of vector index types that scale from memory to disk, and it can fuse vector similarity, full-text search, and structured filters into a single query.
It writes ahead to a log for durability, allows multiple processes to read the same collection concurrently, and includes zvec-grep, a local-first workspace search tool that combines pattern search, keyword ranking, and vector search behind one CLI for humans and AI agents. It has been used in production within Alibaba Group.
It is written in C++ and released under the Apache-2.0 license.