Launch your AI SaaS in days, not months.

Production-ready Next.js starter kits with auth, Stripe billing, multi-tenancy, admin tools, and AI wired to 19 providers. Skip setup and start selling faster.

153 tests. Pay once, own it forever.

Three months in. Zero users. Still writing auth.

Every AI SaaS starts the same way: weeks of plumbing before a single user signs up.

Auth & accounts

Sign-up, OAuth, email verification, password resets, sessions. Weeks of work your users never thank you for.

Stripe billing

Checkout, webhooks, plan upgrades, the billing portal. One missed webhook away from losing revenue.

Multi-tenancy

Workspaces, member roles, invitations, permission checks everywhere. The thing you only get right on your third SaaS.

Admin & ops

User management, usage limits, analytics, feature flags. All before your first customer exists.

StackBlaze ships all of it. Week one is product week.

Three ways to build a foundation

Hand-build it

10–12 weeks

Code you write, test, debug, and maintain. Before your product work even starts.

Prompt an agent

Days of babysitting

Generated code you still need to read, understand, and fix at 2am.

StackBlaze kit

Shipping by Friday

Human-written. 153 automated tests. Every edge case handled. One-command install.

Featured template

AI SaaS Kit

A multi-tenant AI SaaS ready for its first customer: auth, workspaces, Stripe billing, usage metering, and streaming chat wired to 19 AI providers.

19 AI providersMulti-tenantStripe billing153 tests

AI Chat, included

AI chat that ships to production, not just localhost.

Built on AI SDK v6 with native streaming, multi-provider support, and full conversation history in PostgreSQL. Everything you need to charge for an AI product, not just demo one.

19 providers, one env var

Switch from OpenAI to Claude to Gemini to Grok with a single config change. Zero code rewrites. Every provider has a sensible default model pre-mapped.

Native streaming via AI SDK v6

Token-by-token streaming with streamText() and toUIMessageStreamResponse(). Not polling, not batching. Real SSE the way users expect.

PostgreSQL conversation history

Full history stored in Postgres with Drizzle ORM. Conversations and messages are org-scoped and indexed. Scales with your database, not a third-party limit.

Built-in tools + agentic loops

Calculator, web search (Tavily / Exa / Firecrawl), and SSRF-protected URL fetching pre-wired. Up to 5 tool-call steps per response. Add your own in /lib/ai/tools.

One-command Postgres

docker-compose.yml ships with the kit: Postgres 16, named volume, health check, ready in seconds. Same file works locally and on any Docker host in production.

Deploy anywhere Postgres runs

Vercel + Neon, Railway, AWS RDS, Supabase, GCP Cloud SQL, Azure Database. Swap the DATABASE_URL and ship.

The price isn't the cost. The build is.

Every week you spend on auth is a week your competitor spent on their product. The AI SaaS Kit replaces roughly 10–12 weeks of senior-engineering effort: auth, billing, multi-tenancy, admin. None of which ships your product.

Clone it. Add your keys. Deploy.

Pay once. Yours forever. No subscriptions, no seat licenses, no nonsense.