Production-ready full-stack AI SaaS boilerplate with multi-tenancy, subscription billing, and GDPR compliance.

A complete starter template for building AI-powered SaaS applications. Covers the full stack from authentication and billing to LLM integration and compliance — the kind of plumbing every SaaS needs but nobody wants to build twice.

What makes this interesting:

• Custom LLM abstraction layer that routes between OpenAI and Anthropic with automatic fallback
• Row-level multi-tenancy with tenant isolation enforced at every data access point
• Every AI call is metered (tokens, cost, latency, provider) for usage-based billing
• Stripe subscription lifecycle fully managed via webhooks
• GDPR export, deletion, and audit log anonymization built in from day one

graph LR
    Browser --> NextJS["Next.js 14"]
    NextJS -->|"/api/* proxy"| FastAPI["FastAPI"]
    FastAPI --> PostgreSQL["PostgreSQL"]
    FastAPI --> AI["AI Module"]
    FastAPI --> Stripe["Stripe"]
    FastAPI --> GDPR["GDPR Service"]
    AI --> OpenAI["OpenAI"]
    AI --> Anthropic["Anthropic"]
    Stripe -->|Webhooks| FastAPI

AI Module — Chat, summarize, analyze with pluggable LLM abstraction (OpenAI, Anthropic, custom providers)

Multi-tenancy — Row-level data isolation between tenants with tenant_id on every scoped table

Auth & RBAC — JWT access + refresh tokens, bcrypt hashing, roles: owner / admin / member

Stripe Billing — Three-tier subscriptions (free / starter / pro), webhook handlers, billing portal

GDPR Compliance — Tenant & user data export, deletion (Art. 17), audit log anonymization (Art. 20)

Production Ready — Docker Compose, GitHub Actions CI, 93% test coverage, structured logging

git clone https://github.com/ForwardCodeSolutions/ai-saas-starter.git
cd ai-saas-starter
cp .env.example .env    # edit with your API keys
docker compose up -d

Prerequisites: Python 3.11+, Node.js 20+, PostgreSQL, uv

# Backend
cp .env.example .env          # configure DATABASE_URL and API keys
uv sync                       # install Python dependencies
cd backend && uv run alembic upgrade head && cd ..   # run migrations
make dev                      # or: uv run uvicorn backend.src.saas_starter.main:app --port 8003

# Frontend
cd frontend
npm install
npm run dev                   # http://localhost:3000

69 tests across unit and integration suites with 93% code coverage.

Unit tests:     31 tests (models, security, auth, LLM router, Stripe, GDPR service)
Integration:    38 tests (auth API, AI API, billing API, GDPR API, full SaaS flow)

make test          # run all tests with coverage
make lint          # ruff check + format
make check         # lint + test (run before every commit)

├── backend/src/saas_starter/
│   ├── ai/                    # LLM abstraction layer
│   │   ├── providers/         # OpenAI, Anthropic, Mock
│   │   ├── features/          # chat, summarize, analyze
│   │   └── router.py          # model-prefix routing + fallback
│   ├── api/v1/                # FastAPI routers
│   ├── core/                  # config, database, security, exceptions
│   ├── models/                # SQLAlchemy ORM models
│   ├── schemas/               # Pydantic request/response
│   └── services/              # auth, stripe, gdpr, usage
├── frontend/
│   ├── app/                   # Next.js 14 app router pages
│   ├── components/            # Navbar, AuthGuard
│   └── lib/                   # API client (axios)
├── tests/
│   ├── unit/                  # 6 unit test modules
│   └── integration/           # 5 integration test modules
├── docs/
│   ├── decisions/             # 5 Architecture Decision Records
│   ├── api.md                 # API endpoint reference
│   ├── architecture.md        # System design overview
│   ├── code-conventions.md    # Code style guide
│   └── testing-strategy.md    # Testing approach
├── docker-compose.yml
├── Makefile
└── pyproject.toml

This is a starter template. Review and update security settings, secret keys, and CORS configuration before production use.

MIT © 2026 ForwardCodeSolutions