OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
RAG & Knowledge Systems

Wundeedb.jl

WunDeeDB.jl is an embedded Julia vector database library that stores and queries numerical embeddings in a SQLite database, supporting keyed retrieval and distance search with a linear fallback or approximate-neighbor indexes such as HNSW and LM-DiskANN.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Wundeedb.jl — OpenViST project profile in RAG & Knowledge Systems.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when a Julia application needs SQLite-backed embedding storage and vector search as a component, rather than a complete document-retrieval or generation application.

Who it is for

  • Julia developers building applications that need embedded vector storage.
  • Teams developing retrieval-oriented applications that want a SQLite-backed component.

Documented uses

  • Store keyed embedding vectors in a SQLite database.
  • Retrieve one or multiple embeddings by key, and delete entries by key.
  • Search for nearest vectors using linear distance queries or an enabled ANN index.
Before you choose

Requirements & limits

The repository is public and identifies the project as MIT-licensed. Its README provides a Julia package-manager installation instruction.

  • The package is written in Julia and uses SQLite as its backend.
  • The README demonstrates Float16, Float32, Float64 and integer numeric types.
  • Linear distance search is available as the documented fallback when no ANN method is enabled.
  • The README shows HNSW and LM-DiskANN as approximate-neighbor indexing options.
  • The project describes itself as embedded and WAL-based.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Create a database and search embeddings

Set up a SQLite-backed vector database, add sample vectors and run a nearest-neighbor search.

Before you start

  • Julia package manager, with the README installation example using Julia v1.9
  • WunDeeDB.jl installed from its GitHub repository

Steps

  1. Add the package with (@v1.9) pkg> add https://github.com/mantzaris/WunDeeDB.jl.

  2. In Julia, load WunDeeDB and initialize a database path with a dimension, numeric type and search option, such as initialize_db("demo.sqlite", 3, "Float32"; ann="") for the documented linear fallback.

  3. Insert keyed vectors with insert_embeddings, then call search_ann with a query vector, distance metric and top_k value; the README demonstrates Euclidean distance.

Expected outcome

A SQLite database containing keyed embeddings, with a nearest-neighbor search result from the selected approach.

Guide limitations

  • The README's ANN examples use HNSW or LM-DiskANN; the guide's example selects linear fallback instead.
  • The example uses three-dimensional Float32 vectors; choose dimensions and types appropriate to the application.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
17 October 2026
Original website title
mantzaris/WunDeeDB.jl · Official homepage title · retained in its source language
How we review sourcesSuggest a correction