vectordbz/vectordbz

A modern desktop application for exploring, managing, and analyzing vector databases

Summary

VectorDBZ is an open-source desktop client for working with vector databases, written in TypeScript and released under the MIT license. It connects to local or cloud instances of multiple vector stores, letting engineers browse collections, run vector and hybrid searches, and visualize embeddings in 2D and 3D without standing up any additional infrastructure.

What it is useful for

VectorDBZ is useful when engineers need a GUI to inspect, query, and debug vector databases during development of retrieval-augmented generation (RAG) systems, semantic search features, or embedding-based applications. Instead of writing ad-hoc scripts to peek inside a collection, an engineer can point the app at an instance and interact with it. According to the README, the app supports Qdrant (v1.7+), Weaviate (v1.19+), Milvus (v2.3+), ChromaDB (v0.4+), Pinecone, pgvector on PostgreSQL 11+, Elasticsearch (v8.x), and RedisSearch on Redis Stack (v2.0+). It is also useful for visually comparing embedding spaces across databases, since the same client handles several backends.

How engineers can use it

For most engineers, the fastest path is installing the prebuilt binary. On macOS, the README documents a Homebrew tap:

brew tap vectordbz/vectordbz
brew install --cask vectordbz

On Windows, Linux, and other macOS variants, the README links to a downloads table with .exe installers (Windows 10+), darwin-x64 and darwin-arm64 zips (macOS 10.15+), and .deb/.rpm packages (Ubuntu 18.04+, Fedora 32+). All of these point to https://github.com/vectordbz/vectordbz/releases/latest.

The README explicitly calls out a documented limitation: the macOS build is not code-signed. On first launch the user is told to right-click the app and choose Open, and if macOS reports that the app is damaged, the documented workaround is xattr -cr /Applications/VectorDBZ.app.

For engineers who want to modify or run the app from source, the README points to docs/DEVELOPMENT.md for full prerequisites and scripts. The quick start it lists is:

git clone https://github.com/vectordbz/vectordbz.git
cd vectordbz/app
npm ci
npm run start

If the development prerequisites or test data setup are unclear, the README directs them to docs/DEVELOPMENT.md rather than describing them inline, so that file is the authoritative source for local development.

Contributions are accepted via CONTRIBUTING.md, and the README links to a step-by-step guide, docs/ADDING_A_DATABASE.md, for engineers who want to add support for another vector backend.