This is the 2nd launch from Semafind Knotes. View more

Firebase Vector Search

Firebase vector search made easy

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.

Free

Launch tags:

Developer Tools•Artificial Intelligence•GitHub

Launch Team

nuric

Wishoom

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 🚀

Report

1yr ago

nuric

Wishoom

Maker

@mapnamer Thank you for your comment and support! Hybrid search is definitely on the roadmap.

Report

1yr ago

nuric

Wishoom

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.

Report

1yr ago