supabase-reference-architecture
Implement Supabase reference architecture with best-practice project layout. Use when designing new Supabase integrations, reviewing project structure, or establishing architecture standards for Supabase applications. Trigger with phrases like "supabase architecture", "supabase best practices", "supabase project structure", "how to organize supabase", "supabase layout".
Install
mkdir -p .claude/skills/supabase-reference-architecture && curl -L -o skill.zip "https://mcp.directory/api/skills/download/8586" && unzip -o skill.zip -d .claude/skills/supabase-reference-architecture && rm skill.zipInstalls to .claude/skills/supabase-reference-architecture
About this skill
Supabase Reference Architecture
Overview
Production Supabase applications need more than a flat lib/supabase.ts file. This skill covers five enterprise architecture patterns: monorepo with shared types, multi-tenant RLS isolation, microservices with separate Supabase projects, framework integration (Next.js / SvelteKit), and operational patterns (edge functions, caching, queues, audit trails). Each pattern stands alone — pick the ones that match your scale.
For the full monorepo directory layout and microservices cross-project access, see Project Structure. For edge functions, caching, queue, and audit trail patterns, see Operational Patterns.
Prerequisites
@supabase/supabase-jsv2+ installed (npm install @supabase/supabase-js)- Supabase CLI installed (
npm install -g supabase) - A Supabase project at supabase.com/dashboard
- Familiarity with
supabase-install-auth(project URL, anon key, service role key) - PostgreSQL basics (RLS policies, triggers, functions)
Instructions
Step 1: Client Singleton — The Foundation
Every app in the monorepo imports from a shared package instead of creating its own client. This guarantees a single source of truth for the URL, keys, and type definitions.
// packages/supabase/src/client.ts
import { createClient, SupabaseClient } from '@supabase/supabase-js'
import type { Database } from './database.types'
let client: SupabaseClient<Database> | null = null
export function getSupabaseClient(): SupabaseClient<Database> {
if (!client) {
const url = process.env.NEXT_PUBLIC_SUPABASE_URL ?? process.env.SUPABASE_URL
const key = process.env.NEXT_PUBLIC_SUPABASE_ANON_KEY ?? process.env.SUPABASE_ANON_KEY
if (!url || !key) {
throw new Error('Missing SUPABASE_URL or SUPABASE_ANON_KEY environment variables')
}
client = createClient<Database>(url, key)
}
return client
}
// Reset for testing
export function resetClient(): void {
client = null
}
// packages/supabase/src/admin.ts — Server-side only, never bundle in client code
import { createClient } from '@supabase/supabase-js'
import type { Database } from './database.types'
export function getSupabaseAdmin() {
const url = process.env.SUPABASE_URL
const serviceKey = process.env.SUPABASE_SERVICE_ROLE_KEY
if (!url || !serviceKey) {
throw new Error('Missing SUPABASE_URL or SUPABASE_SERVICE_ROLE_KEY — server-only')
}
return createClient<Database>(url, serviceKey, {
auth: { autoRefreshToken: false, persistSession: false }
})
}
Key detail: The admin client sets autoRefreshToken: false and persistSession: false because server-side code should never store user sessions.
Step 2: Multi-Tenant RLS via JWT Claims
The most scalable Supabase multi-tenant pattern uses a custom JWT claim (org_id) combined with RLS policies. Every table includes an org_id column, and RLS extracts the tenant from the user's JWT — no application-level filtering needed.
-- Migration: 20260101000000_create_tenants.sql
-- Tenants table
create table public.tenants (
id uuid primary key default gen_random_uuid(),
name text not null,
slug text unique not null,
plan text default 'free' check (plan in ('free', 'pro', 'enterprise')),
created_at timestamptz default now()
);
-- Tenant membership
create table public.tenant_members (
tenant_id uuid references public.tenants(id) on delete cascade,
user_id uuid references auth.users(id) on delete cascade,
role text default 'member' check (role in ('owner', 'admin', 'member', 'viewer')),
primary key (tenant_id, user_id)
);
-- Example tenant-scoped table
create table public.projects (
id uuid primary key default gen_random_uuid(),
org_id uuid not null references public.tenants(id) on delete cascade,
name text not null,
created_by uuid references auth.users(id),
created_at timestamptz default now()
);
-- Enable RLS on all tenant-scoped tables
alter table public.projects enable row level security;
-- RLS policy: users can only see rows belonging to their tenant
-- The org_id is extracted from the JWT claims set during authentication
create policy "Tenant isolation" on public.projects
for all
using (
org_id = (auth.jwt() ->> 'org_id')::uuid
);
The tenant-switching function verifies membership before updating the JWT claim:
-- Helper function to set org_id in JWT claims after login
create or replace function public.set_tenant_claim(tenant_id uuid)
returns void as $$
begin
-- Verify user is a member of this tenant
if not exists (
select 1 from public.tenant_members
where tenant_members.tenant_id = set_tenant_claim.tenant_id
and tenant_members.user_id = auth.uid()
) then
raise exception 'Not a member of tenant %', tenant_id;
end if;
-- Set the custom claim
perform auth.update_user_metadata(
auth.uid(),
jsonb_build_object('org_id', tenant_id)
);
end;
$$ language plpgsql security definer;
Key details for multi-tenant RLS:
auth.jwt() ->> 'org_id'reads a custom claim from the user's JWT — zero application code needed- Every tenant-scoped table must have an
org_idcolumn and RLS enabled - Tenant switching requires updating the JWT claim and re-authenticating
- For row-level tenant + role permissions, combine
org_idwith a role lookup
Step 3: Framework Integration (Next.js)
Server components use the service_role key for direct database access. Client components use the anon key with RLS protection.
// app/lib/supabase-server.ts — Next.js App Router (server components)
import { createClient } from '@supabase/supabase-js'
import { cookies } from 'next/headers'
import type { Database } from '@my-platform/supabase'
export async function getSupabaseServer() {
const cookieStore = await cookies()
return createClient<Database>(
process.env.NEXT_PUBLIC_SUPABASE_URL!,
process.env.SUPABASE_SERVICE_ROLE_KEY!,
{
auth: { autoRefreshToken: false, persistSession: false },
global: {
headers: {
// Forward the user's auth cookie for RLS context
cookie: cookieStore.toString()
}
}
}
)
}
// app/projects/page.tsx — Server component with direct DB access
export default async function ProjectsPage() {
const supabase = await getSupabaseServer()
const { data: projects } = await supabase
.from('projects')
.select('id, name, created_at')
.order('created_at', { ascending: false })
.limit(50)
return <ProjectList projects={projects ?? []} />
}
// app/lib/supabase-browser.ts — Client components use the anon key
'use client'
import { createClient } from '@supabase/supabase-js'
import type { Database } from '@my-platform/supabase'
let browserClient: ReturnType<typeof createClient<Database>> | null = null
export function getSupabaseBrowser() {
if (!browserClient) {
browserClient = createClient<Database>(
process.env.NEXT_PUBLIC_SUPABASE_URL!,
process.env.NEXT_PUBLIC_SUPABASE_ANON_KEY!
)
}
return browserClient
}
For SvelteKit integration and additional framework patterns, see Examples.
Output
After applying these patterns you will have:
- Monorepo with shared Supabase client, typed database access, and centralized migrations
- Multi-tenant RLS isolation using
auth.jwt() ->> 'org_id'— zero application-level filtering - Framework-specific integration for Next.js (server/client split) and SvelteKit (hooks)
- Edge Functions, caching layer, job queue, and audit trail (see Operational Patterns)
Error Handling
| Error | Cause | Solution |
|---|---|---|
Missing SUPABASE_URL or SUPABASE_ANON_KEY | Environment variables not set | Check .env file and ensure variables are loaded |
new row violates row-level security policy | RLS blocks the operation | Verify org_id JWT claim matches the row's org_id |
Not a member of tenant | User tried switching to unauthorized tenant | Check tenant_members table for the user-tenant pair |
TypeError: Cannot read properties of null | Client singleton not initialized | Ensure env vars are available before first getSupabaseClient() call |
cron.schedule: permission denied | pg_cron extension not enabled | Enable via dashboard: Database > Extensions > pg_cron |
For the full error reference including RLS debugging and cross-project troubleshooting, see Error Handling Reference.
Examples
Multi-Tenant Query Flow (TypeScript)
import { createClient } from '@supabase/supabase-js'
import type { Database } from './database.types'
const supabase = createClient<Database>(
process.env.SUPABASE_URL!,
process.env.SUPABASE_ANON_KEY!
)
// 1. Sign in
const { data: { session } } = await supabase.auth.signInWithPassword({
email: 'user@example.com',
password: 'secure-password'
})
// 2. Switch tenant context
const { error: claimError } = await supabase.rpc('set_tenant_claim', {
tenant_id: 'tenant-uuid-here'
})
if (claimError) throw claimError
// 3. Refresh session to pick up new JWT claims
await supabase.auth.refreshSession()
// 4. All subsequent queries are automatically scoped to this tenant
const { data: projects } = await supabase
.from('projects')
.select('id, name, created_at')
.order('created_at', { ascending: false })
console.log('Tenant projects:', projects)
// Only returns projects where org_id matches the JWT claim
For the job queue consumer example and SvelteKit integration, see Examples.
Resources
- Supabase Architecture
- Row Level Security
- [Multi-Tenant RLS](https://supabase.com/docs/guide
Content truncated.
More by jeremylongshore
View all skills by jeremylongshore →You might also like
flutter-development
aj-geddes
Build beautiful cross-platform mobile apps with Flutter and Dart. Covers widgets, state management with Provider/BLoC, navigation, API integration, and material design.
drawio-diagrams-enhanced
jgtolentino
Create professional draw.io (diagrams.net) diagrams in XML format (.drawio files) with integrated PMP/PMBOK methodologies, extensive visual asset libraries, and industry-standard professional templates. Use this skill when users ask to create flowcharts, swimlane diagrams, cross-functional flowcharts, org charts, network diagrams, UML diagrams, BPMN, project management diagrams (WBS, Gantt, PERT, RACI), risk matrices, stakeholder maps, or any other visual diagram in draw.io format. This skill includes access to custom shape libraries for icons, clipart, and professional symbols.
ui-ux-pro-max
nextlevelbuilder
"UI/UX design intelligence. 50 styles, 21 palettes, 50 font pairings, 20 charts, 8 stacks (React, Next.js, Vue, Svelte, SwiftUI, React Native, Flutter, Tailwind). Actions: plan, build, create, design, implement, review, fix, improve, optimize, enhance, refactor, check UI/UX code. Projects: website, landing page, dashboard, admin panel, e-commerce, SaaS, portfolio, blog, mobile app, .html, .tsx, .vue, .svelte. Elements: button, modal, navbar, sidebar, card, table, form, chart. Styles: glassmorphism, claymorphism, minimalism, brutalism, neumorphism, bento grid, dark mode, responsive, skeuomorphism, flat design. Topics: color palette, accessibility, animation, layout, typography, font pairing, spacing, hover, shadow, gradient."
godot
bfollington
This skill should be used when working on Godot Engine projects. It provides specialized knowledge of Godot's file formats (.gd, .tscn, .tres), architecture patterns (component-based, signal-driven, resource-based), common pitfalls, validation tools, code templates, and CLI workflows. The `godot` command is available for running the game, validating scripts, importing resources, and exporting builds. Use this skill for tasks involving Godot game development, debugging scene/resource files, implementing game systems, or creating new Godot components.
nano-banana-pro
garg-aayush
Generate and edit images using Google's Nano Banana Pro (Gemini 3 Pro Image) API. Use when the user asks to generate, create, edit, modify, change, alter, or update images. Also use when user references an existing image file and asks to modify it in any way (e.g., "modify this image", "change the background", "replace X with Y"). Supports both text-to-image generation and image-to-image editing with configurable resolution (1K default, 2K, or 4K for high resolution). DO NOT read the image file first - use this skill directly with the --input-image parameter.
pdf-to-markdown
aliceisjustplaying
Convert entire PDF documents to clean, structured Markdown for full context loading. Use this skill when the user wants to extract ALL text from a PDF into context (not grep/search), when discussing or analyzing PDF content in full, when the user mentions "load the whole PDF", "bring the PDF into context", "read the entire PDF", or when partial extraction/grepping would miss important context. This is the preferred method for PDF text extraction over page-by-page or grep approaches.
Related MCP Servers
Browse all serversGitHub Chat lets you query, analyze, and explore GitHub repositories with AI-powered insights, understanding codebases f
Nekzus Utility Server offers modular TypeScript tools for datetime, cards, and schema conversion with stdio transport co
Break down complex problems with Sequential Thinking, a structured tool and step by step math solver for dynamic, reflec
Build persistent semantic networks for enterprise & engineering data management. Enable data persistence and memory acro
Boost your AI code assistant with Context7: inject real-time API documentation from OpenAPI specification sources into y
Boost productivity with Task Master: an AI-powered tool for project management and agile development workflows, integrat
Stay ahead of the MCP ecosystem
Get weekly updates on new skills and servers.