Ship your AI SaaSin a weekend

Auth, Postgres, Stripe subscriptions, transactional email and a streaming AI layer — wired together and ready to deploy. MIT licensed, so whatever you build on it is yours.

John D.
Alice S.
David J.
Michael W.
Sophia G.

Trusted by 10,000+ forward-thinking teams

Dashboard preview

Built on technology you trust

Features

Everything a working AI product needs.

Streaming chat, one function call

Persisted conversations, file attachments and a model picker across OpenAI and Anthropic. Rate limited before a single token is spent.

route.ts
// app/api/chat/route.ts
 
const result = await streamChat({
messages,
conversationId,
modelId,
});
 
// usage is recorded once the stream ends
await updateUsage({ teamId, totalTokens });

Structured output

Task templates force the model to return JSON that matches a Zod schema — validated before it reaches your UI.

Subscriptions

Three Stripe tiers with Checkout, the customer portal, and webhooks that keep plan state in your own database.

3 weeks

Of auth, billing and AI plumbing you don't have to write

15 min

From clone to a running app on your own keys

$0

MIT licensed — no seat fees, no revenue share, ever

Usage limits that actually hold

Daily and monthly request caps per plan, enforced server-side before the model is called, with token counts recorded per team so you know what a customer really costs.

In the box

Twelve things you don't have to build.

Every one of these is wired up and working on first run — not a stub, not a TODO.

  • Authentication

    Email/password and Google OAuth with sessions, verification and password reset, powered by Better-Auth.

  • Stripe Billing

    Three subscription tiers wired to Checkout and webhooks, with plan state stored in your database.

  • Customer Portal

    Self-serve billing management, plan upgrades and invoices, straight from Stripe.

  • AI Chat

    Streaming chat on the Vercel AI SDK with persisted history and a model picker.

  • Structured Output

    One-shot task templates that return validated JSON against a Zod schema.

  • Usage Limits

    Per-plan daily and monthly request caps enforced server-side before a token is spent.

  • Admin Dashboard

    Overview of users, workspaces and usage across your whole app.

  • Database ORM

    Drizzle ORM on Postgres with typed schemas and a CLI migration workflow.

  • Email

    Transactional email through Resend for verification and account flows.

  • MDX Blog & Docs

    Contentlayer-powered blog and help centre so your marketing site ships with the app.

  • Dark Mode

    Light and dark themes built on CSS variables and shadcn/ui.

  • MIT Licensed

    Fork it, sell what you build with it, no strings attached.

How it works

From clone to deployed in three steps.

  1. 01

    Clone

    Pull the repo, install, and copy the example environment file.

    $npx degit jeanchrinot/shipaisaas-starter my-app && npm install
  2. 02

    Configure

    Add your Postgres URL and your OpenAI, Resend and Stripe keys, then create the schema.

    $cp .env.example .env && npm run db:migrate
  3. 03

    Ship

    Edit the system prompt, add your own task templates, and deploy.

    $npm run dev → vercel deploy

Loved by builders

Teams that ship faster with AI agents.

See how operators and founders are automating their workflows and reclaiming time for the work that actually moves the needle.

The Ship AI SaaS boilerplate has truly revolutionized my workflow. The decoupled orchestrator and built-in QStash background queues saved me months of architecture design. I shipped my vertical AI product in a single weekend.
John Doe

John Doe

Full Stack Developer

I wanted to build an AI tool for legal tech but dreaded the Stripe and Auth setup. This starter kit gave me the plumbing and the LLM routing out of the box. I just focused on my custom agent prompts and hit deploy.
Alice Smith

Alice Smith

SaaS Founder

Handling long-running agent tasks without Vercel timeouts was a nightmare until I used this boilerplate. The serverless-proof queueing is brilliantly implemented.
David Johnson

David Johnson

DevOps Engineer

The 'Human-in-the-Loop' guardrails are exactly what our enterprise clients demand. Being able to pause an agent, review its work, and approve it directly in the chat UI is an absolute game-changer for safety.
Michael Wilson

Michael Wilson

Product Manager

I love how easy it is to add a new tool to the registry. The UI and LLM routers instantly pick it up. Plus, the Zod-enforced schemas make it mathematically impossible for the LLM to hallucinate function arguments.
Sophia Garcia

Sophia Garcia

AI Engineer

I skipped all the boring stuff. Auth, the Stripe customer portal, and the dashboard were ready on day one. I was able to start charging customers immediately while I perfected my multi-agent workflows.
Emily Brown

Emily Brown

Indie Hacker

The Chat-as-State UI is a masterpiece. Users can actually watch the agents 'think' and execute step-by-step in the background ledger. It builds so much trust with the end user compared to a loading spinner.
Jason Stan

Jason Stan

Software Architect

The Ship AI SaaS boilerplate has truly revolutionized my workflow. The decoupled orchestrator and built-in QStash background queues saved me months of architecture design. I shipped my vertical AI product in a single weekend.
John Doe

John Doe

Full Stack Developer

I wanted to build an AI tool for legal tech but dreaded the Stripe and Auth setup. This starter kit gave me the plumbing and the LLM routing out of the box. I just focused on my custom agent prompts and hit deploy.
Alice Smith

Alice Smith

SaaS Founder

Handling long-running agent tasks without Vercel timeouts was a nightmare until I used this boilerplate. The serverless-proof queueing is brilliantly implemented.
David Johnson

David Johnson

DevOps Engineer

The 'Human-in-the-Loop' guardrails are exactly what our enterprise clients demand. Being able to pause an agent, review its work, and approve it directly in the chat UI is an absolute game-changer for safety.
Michael Wilson

Michael Wilson

Product Manager

I love how easy it is to add a new tool to the registry. The UI and LLM routers instantly pick it up. Plus, the Zod-enforced schemas make it mathematically impossible for the LLM to hallucinate function arguments.
Sophia Garcia

Sophia Garcia

AI Engineer

I skipped all the boring stuff. Auth, the Stripe customer portal, and the dashboard were ready on day one. I was able to start charging customers immediately while I perfected my multi-agent workflows.
Emily Brown

Emily Brown

Indie Hacker

The Chat-as-State UI is a masterpiece. Users can actually watch the agents 'think' and execute step-by-step in the background ledger. It builds so much trust with the end user compared to a loading spinner.
Jason Stan

Jason Stan

Software Architect

Pricing

Start at full speed.

Free

$0/mo

Kick the tires

  • 50 AI requests per month
  • Chat with history
  • All task templates
  • Basic usage analytics
  • No priority support
  • Shorter context per request
  • No custom branding

Pro

Most popular
$19$15/mo

$180 billed annually

For solo builders shipping real work

  • 1,000 AI requests per month
  • Larger context per request
  • All task templates
  • Full usage analytics
  • Email support
  • No custom branding

Business

$49$39/mo

$468 billed annually

For teams with heavier usage

  • 5,000 AI requests per month
  • Largest context per request
  • All task templates
  • Full usage analytics
  • Priority support

Questions? support@shipaisaas.com · Test cards won't be charged.

Ready to ship your AI SaaS?

Clone the starter, add your keys, and have a working product with auth, billing and AI in an afternoon.

Start for $0

FAQ

Common questions, straight answers.