supabaseVectorStore

import { supabaseVectorStore } from '@agentskit/memory'

const store = supabaseVectorStore({
  url: process.env.SUPABASE_URL!,
  serviceRoleKey: process.env.SUPABASE_SERVICE_ROLE_KEY!,
})

@supabase/supabase-js is an optional peer dependency loaded lazily.

#Server-side setup

Run once in the Supabase SQL editor:

create extension if not exists vector;

create table agentskit_vectors (
  id text primary key,
  content text,
  embedding vector(1536),
  metadata jsonb
);

create or replace function agentskit_execute_sql(sql text, params jsonb)
  returns setof json language plpgsql security definer as $$
begin
  return query execute sql using params;
end;
$$;

The wrapper expects the agentskit_execute_sql RPC to exist — it's how we channel parameterized SQL from the Supabase client to pgvector without bundling a separate pg driver.

#Config

#Surface

Implements VectorMemory: store(docs) / search(embedding, options) / delete(ids). Inherits the VectorFilter contract — pass options.filter for metadata-scoped retrieval.

#Caveats

• Service-role key bypasses RLS — keep it server-side only.
• Cosine-distance search is the default; switch operators in the underlying RPC if you need a different metric.

#Related

• pgvector — the adapter this wraps.
• Memory overview