Weaviate DB

The AI-Native Vector Database Workbench for VS Code

Inspect schemas · Query objects · Import data · Run GraphQL — without leaving your editor.

Overview

Weaviate DB is a zero-config GUI workbench for Weaviate — the leading AI-native vector database. Connect to any local or cloud Weaviate instance in seconds and get a full-featured panel for schema exploration, object browsing, data imports, and GraphQL querying — all inside your editor.

No browser tabs. No separate apps. Just your editor.

Demo

💻 Local Instance

🎬 Watch the Intro Video

Features

Quick Start

Step 1 — Install

Search Weaviate DB in the Extensions panel (Ctrl+Shift+X) or run:

ext install harishkaparwan.weaviate-db

Step 2 — Start Weaviate (skip if already running)

docker run --rm -p 8083:8080 -p 50051:50051 \
  docker.io/semitechnologies/weaviate:1.37.4

Step 3 — Open the Workbench

Click the Weaviate DB icon in the Activity Bar, or open the Command Palette and run:

Weaviate DB: Open Workbench

Step 4 — Connect

Your schema, objects, and query interface load instantly. ✅

Sample Data for Import

Use the Insert tab to import data from a file (JSON or CSV) or a direct URL.
The datasets below are ready to paste into the Import from URL field or download and drag in as a file.

🧠 Pre-vectorized Data — Bring Your Own Vectors

These files already contain embedding vectors. Use them when you want full control over the embedding model or are migrating from another system.

How to import a pre-vectorized file:

1. Open the Insert tab → set Mode to Batch objects
2. Paste the direct JSON URL into Import from URL and click Import URL, or click Import file and select the downloaded file
3. Click Insert — vectors are stored as-is, no re-embedding happens

⚙️ Non-vectorized Data — Vectorized at Import Time

These files contain only raw text / structured fields. Weaviate (or an external module) generates the vectors during import based on your class vectorizer configuration.

How to import a non-vectorized file:

1. Open the Insert tab → set Mode to Batch objects
2. Make sure your Weaviate class has a vectorizer configured (e.g. text2vec-openai)
3. Paste a direct URL or import the file — Weaviate calls the vectorizer module automatically at insert time

📐 Index Type Reference

You can inspect and change a class's vector index type from the Index tab in Weaviate DB after connecting.

Connection Modes

🖥️ Local Mode

Connect to any Weaviate instance running on your machine or local network. No authentication headers are added.

http://localhost:8083

Works with Docker, local binaries, or a proxy like http://localhost:8787.

☁️ Cluster Mode (Weaviate Cloud)

Enter your cluster REST host and API key. Authentication is handled automatically.

https://<cluster-id>.c0.<region>.gcp.weaviate.cloud

Tip: Use the REST endpoint — not the Weaviate Cloud Console URL.

NPM Package

You can also run Weaviate UI directly from NPM without any installation:

# Install globally
npm install -g weaviate-db-extension

# Run directly with npx
npx weaviate-db-extension

NPM Package: weaviate-db-extension

Available On Every Platform

Homebrew (v2.1.10)

brew tap harishkaparwan/tap
brew install --cask weaviate-ui

Latest Version: 2.1.10 (includes Windows desktop support) Platforms: macOS Apple Silicon, Windows EXE/MSI, Web, NPM Features: Cross-platform desktop apps with native installers

Docker

docker run --rm -p 8080:80 harishkaparwan/weaviate-ui:latest

Works With Every VS Code-Compatible Editor

Any editor that supports the VS Code Extension API can run Weaviate DB out of the box.

Requirements

• VS Code ^1.85.0 or any compatible editor
• A running Weaviate instance (local or cloud)
• No additional dependencies or configuration files needed

Support & Feedback