Why you should use Edge Functions for AI

Edge Functions are perfect for encapsulating AI logic, but downloading these massive models on every invocation leads to significant load times, and a very large bill! In this video, Jon Meyers explains the benefits of Supabase attaching these models on disk.

00:00 Don't Download AI Models in Edge Functions
00:24 Overview of AI Inference example
00:48 Create a vector embedding
02:08 Performing similarity search in Edge Function
03:06 Sending a HTTP request to search embeddings
03:41 This entire system running on LocalHost

💻 Videos to watch next:

👇 Learn more about Supabase 👇

📱 Connect with Us:

ABOUT SUPABASE:
Supabase is the open source Firebase alternative. Supabase provides a full Postgres database for every project with pgvector, backups, realtime, and more. Add and manage email and password, passwordless, OAuth, and mobile logins to your project through a suite of identity providers and APIs.

Build in a weekend, scale to millions.

#Supabase #AppDevelopment #RealtimeApps #DeveloperTools

[Supabase]

If you wanna go deeper, check out the blog 👇

[sadmanyasar]

If anyone was confused about creating the database function like me, you need to create a webhook first using the generate-embedding edge function then pass your anon key from project settings in the authorization header value with the initial "Bearing " e.g Authorization: "Bearing YOUR_ANON_KEY". Then you will find the webhook is added in the database trigger automatically as shown in 1:13. Also make sure to enable Enforce JWT Verification in the edge function.

[emrememis]

This is very helpful tutorial, however, I see one potential problem in your codebase. You have enabled the webhook for inserting and updating, and then you have updated the table in the edge-function. Can there be an infinite loop?

[yourpc]

how to fix "permission denied for schema public" in edge supabase?

[dshukertjr]

A controvertial way of pronouncing tomato 🍅