Multi-Tenancy

Built for Multi-Tenancy From Day One

Row-level security, workspace isolation, per-tenant configuration, resource quotas, and white-label support. Multi-tenancy is not bolted on — it is the foundation.

Data Isolation

Three Layers of Tenant Isolation

Multi-tenancy is not an afterthought — isolation is enforced at the database, application, and storage layers simultaneously.

Database Layer

Row-level security on every table

PostgreSQL RLS policies on all tenant-scoped tables
Tenant ID injected automatically via session context
No application-level filtering — security enforced at database level
Cross-tenant queries are impossible by design

Application Layer

Request-scoped tenant context

Tenant resolved from API key, JWT, or subdomain
Context propagated through entire request lifecycle
Middleware validates tenant on every request
Agent, memory, and tool access scoped to tenant

Storage Layer

Isolated file and vector storage

S3 prefixes scoped per tenant for file uploads
Vector DB collections isolated per workspace
Redis key namespacing prevents cache collisions
Encryption keys rotatable per tenant
Configuration

Every Tenant Gets Their Own Setup

Different agents, different models, different memory strategies, different branding — all managed from one admin panel.

TenantAgentsLLM ModelMemoryBranding
Salon A3GPT-4oSession + Long-termPink theme, custom logo
Clinic B5Claude 3.5All 5 tiersBlue theme, HIPAA mode
Agency C12Mixed (GPT-4o + Gemini)Session onlyWhite-label, custom domain

Each row is a fully independent workspace — no shared state

Resource Management

Prevent Noisy Neighbors with Quotas

Set granular limits on tokens, API calls, storage, and sessions per tenant. Enforce fairness without manual monitoring.

Token Budget

Maximum tokens per day/month per tenant

Salon A: 500K tokens/month

API Rate Limit

Maximum requests per minute per tenant

Clinic B: 100 req/min

Storage Quota

Maximum file and knowledge base storage

Agency C: 10GB total

Concurrent Sessions

Maximum simultaneous active conversations

Salon A: 20 concurrent

Agent Limit

Maximum number of agents per workspace

Clinic B: 10 agents

Knowledge Base Size

Maximum documents in knowledge base

Agency C: 5,000 documents

Live Tenant Usage

Salon A
62% of 500K tokens
Clinic B
88% of 1M tokens
Agency C
34% of 2M tokens
White-Label

Your Platform, Their Brand

Custom logos, colors, domains, and agent personalities for every tenant — powered by OrchStack under the hood.

Custom Branding

Each tenant gets their own logo, color scheme, fonts, and favicon. The chat widget, dashboard, and emails all reflect the tenant's brand.

Custom Domains

Map tenant-specific subdomains or custom domains to their workspace. SSL certificates are provisioned and renewed automatically.

Brand-Specific Agents

Each tenant's agents use their own personality, tone, and terminology. The same agent type behaves differently for each brand.

Salon A

Powered by OrchStack

Clinic B

Powered by OrchStack

Agency C

Powered by OrchStack

Platform Features

Everything for Multi-Tenant AI

From database-level isolation to white-label branding — every feature is designed for multi-tenant SaaS from the ground up.

Row-Level Security

PostgreSQL RLS policies ensure every database query is automatically scoped to the requesting tenant. Zero shared state, zero data leaks.

Workspace Isolation

Each tenant gets a fully isolated workspace — agents, memory, knowledge bases, tools, and audit logs are all scoped and separated.

Per-Tenant Config

Each tenant can have different agents, different LLM providers, different guardrail rules, and different workflows. No code changes needed.

Resource Quotas

Set per-tenant limits on token usage, API calls, storage, and concurrent sessions. Prevent noisy-neighbor problems before they start.

White-Label

Custom branding per tenant — logos, colors, domain, email templates. Your AI platform, their brand experience.

SSO per Tenant

Each tenant can configure their own SSO provider — SAML, OIDC, Google Workspace, or Azure AD. Managed from the admin panel.

Multi-Tenancy FAQ

Build Multi-Tenant AI Products

Ship AI features to thousands of tenants with isolation, quotas, and branding built in.

Free tier supports multi-tenancy -- scale when ready