# django-semantic-search: semantic search capabilities integrated with Django ORM

**URL:** <https://forum.djangoproject.com/t/django-semantic-search-semantic-search-capabilities-integrated-with-django-orm/34631>\
**Category:** Show & Tell\
**Created:** [September 9, 2024, 12:03pm UTC](https://forum.djangoproject.com/t/django-semantic-search-semantic-search-capabilities-integrated-with-django-orm/34631 "2024-09-09T12:03:40Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![kacperlukawski](https://sea2.discourse-cdn.com/flex026/user_avatar/forum.djangoproject.com/kacperlukawski/32/23744_2.png) [@kacperlukawski](https://forum.djangoproject.com/u/kacperlukawski)\
**Post date:** [September 9, 2024, 12:03pm UTC](https://forum.djangoproject.com/t/django-semantic-search-semantic-search-capabilities-integrated-with-django-orm/34631/1 "2024-09-09T12:03:40Z")

</div>

Hey 👋 I’ve created a semantic search plugin for Django. It integrates with Django ORM with little effort: [django-semantic-search](https://github.com/kacperlukawski/django-semantic-search/). It’s the initial version that supports Sentence Transformers as embedding models and Qdrant as vector store, but the aim is to keep it model and storage agnostic.

[Documentation](https://kacperlukawski.github.io/django-semantic-search/) describes integrating it with an existing Django app. The design was inspired by some well-known packages like `django-import-export`, so everything is configured by creating another class that defines how a model is mapped to a document kept in a vector store.

I feel Django’s built-in search capabilities are rather limited, and semantic search is a great alternative that works even in multilingual or multimodal scenarios (e.g., reverse image search, like Google Lens).

Any feedback would be fantastic!