# WunDeeDB.jl a lightweight sqlite backed vector database

**URL:** <https://discourse.julialang.org/t/wundeedb-jl-a-lightweight-sqlite-backed-vector-database/130151>\
**Category:** Package Announcements\
**Tags:** package, announcement\
**Created:** [June 24, 2025, 1:11am UTC](https://discourse.julialang.org/t/wundeedb-jl-a-lightweight-sqlite-backed-vector-database/130151 "2025-06-24T01:11:34Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![mantzaris](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/mantzaris/32/3852_2.png) [@mantzaris](https://discourse.julialang.org/u/mantzaris)\
**Post date:** [June 24, 2025, 1:11am UTC](https://discourse.julialang.org/t/wundeedb-jl-a-lightweight-sqlite-backed-vector-database/130151/1 "2025-06-24T01:11:34Z")

</div>

`WunDeeDB.jl: An easy to use, zero config, WAL, SQLite backend vector database`

Is now in the general repository, [GitHub - mantzaris/WunDeeDB.jl: Your just-works / zero-config / embedded / WAL: vector database](https://github.com/mantzaris/WunDeeDB.jl)

What WunDeeDB.jl offers:

- Store vector embeddings in a SQLite db with transactions (eg WAL)
- Have the index connectivity indices per node reside within rows tables of the SQLite DB and not rely on them being loaded all into memory (RAM) so larger than memory stores can be used
- Be able to choose between disk based HNSW or LM-DiskANN for the approximate nearest neighbor search (or linear fall back)
- Ability to add or remove vectors after the index has been constructed
- Able to work on little or big endian machines with internal translation mechanism