Firestore Vector Search with SemaDB

Made by semafind

Indexes data from Firebase Firestore into SemaDB for vector search

<10
installs
Works with
Cloud Firestore
Version
0.0.4 | Source code
Tags
search, ai
License
Apache-2.0
Publisher
semafind
Report
Bug
Abuse

How this extension works

Use this extension as a thin wrapper around SemaDB public API. It creates two functions:

  • semadbSync listens to collection document writes in Firestore and if the document contains a vector field, it makes a request to SemaDB to index it. Only the document ID and vector is synced, no other data is sent.
  • semadbSearch makes a vector search request to SemaDB and returns the points.

Additional Setup

Before using this extension you must:

  1. Register on RapidAPI and get an API key for SemaDB.
  2. Create a collection with the desired vector size and distance metric. Take a note of the collection name.

Billing

SemaDB billing and usage limits are available on RapidAPI pricing page.

This extension uses other Firebase or Google Cloud Platform services which may have associated charges:

  • Firestore to write SemaDB point IDs on indexing the document.
  • Cloud Functions to run the aforementioned functions.

When you use Firebase Extensions, you’re only charged for the underlying resources that you use. A paid-tier billing plan is only required if the extension uses a service that requires a paid-tier plan, for example calling to a Google Cloud Platform API or making outbound network requests to non-Google services. All Firebase services offer a free tier of usage. Learn more about Firebase billing.