Subscribe
Sign in
Semafind Knotes
p/semafindOrganise and discover your private knowledge
Start new thread
nuric

Firebase Vector Search - Firebase vector search made easy

by•

SemaDB Firebase extension is a bridge between Firestore and SemaDB to enable easy vector search across documents. It syncs document vectors stored in Firestore and provides a vector search endpoint.

Add a comment

Replies

Best
nuric
Maker
📌
Hello Product Hunt Community 👋 Firebase is great but it doesn’t have in-built search, let alone vector search. With a growing number of potential applications, AI models are becoming more important and require vector based search to enable solutions such as semantic search, retrieval-augmented generation and product recommendations. We are excited to bring an easy-to-use, fully-hosted vector search extension to your Firebase projects. 1. Install with one click from the Firebase Extensions hub. 2. Sync your document vectors to SemaDB automatically. 3. Perform vector search 🔍 with an integrated Firebase cloud function. The extension never shares your data, just the vector representation and stores the Firestore document ID in SemaDB. You can choose which Firestore and SemaDB collections to sync as well as the vector field. To get started, you only need to create a collection in SemaDB cloud with the desired vector size and distance metric, then install the Firebase extension 🚀
nuric
Maker
@mapnamer Thank you for your comment and support! Hybrid search is definitely on the roadmap.
nuric
Maker
@mapnamer Great ideas! SemaDB has a document based schemaless design so going into MySQL, Postgres style relational models are probably not necessary for hybrid search. LanceDB embodies that approach but in terms of fuss, you probably have to install extra dependencies such as pyarrow, save and load from disk, manage indexes etc. On the other hand, SemaDB is a distributed database that automatically indexes, scales and works in a cloud environment with an easy-to-use RESTful API. Developers, for example, do not need to load in pandas data frame on a local machine and upload them to a S3 bucket, setup cloud run instances to serve them etc. Running locally on a single filesystem is much easier - and performant - compared to a cluster, distributed environment and SemaDB is focused on the latter.
Thomas Richard
This is nice and beautiful.
Geofry Okello
Congratulations on the launch, this is awesome and brilliancy of highest level
Jeffrey Bunn
Launching soon!
Awesome. I use Firebase and search is definitely a pain.
Shahmir Faisal
Congrats on the launch Nuric!
Ghost Kitty
Comment Deleted
Celine
I would love to give it a try!