Buy — $99

A ShipFast alternative for AI agent products

Disclosure. We make Agentwright Starter, so read this as a vendor's comparison. Every fact about ShipFast comes from its own public page, linked below, as checked on November 12, 2026. Prices and features change; check the source before you decide.

ShipFast is a Next.js boilerplate for launching a SaaS quickly, and its page says 8,447 makers use it. If your product is an AI agent that customers pay for, the useful question is not which kit is better. It is which parts each one covers, and which parts you would still have to build.

What ShipFast is

ShipFast describes itself as "the NextJS boilerplate with all you need to build your SaaS, AI tool, or any other web app". Its page lists a payment system (Stripe or Lemon Squeezy), a database (MongoDB or Supabase), login with Google OAuth and magic links, emails through Mailgun or Resend, SEO and a blog, UI components and animations, and ChatGPT prompts for writing terms and a privacy policy. The repo comes in JavaScript and TypeScript, for the app router and the pages router.

It is a one-time purchase. Starter is $199 (list $299), All-in is $249 (list $349), and a bundle with the CodeFast course is $299 (list $648). The page says the $100 discount is for the first 8,460 customers. On the Starter card, four lines are greyed out: the Discord community, the leaderboard, the partner discounts and lifetime updates. The All-in card includes them. The FAQ says there is no refund once you have access to the repo.

On AI, the FAQ says ShipFast is built for AI code editors such as Cursor and Copilot: the codebase gives the editor context, so it can write features for you. That is about how you write your app. The page does not list code for running a model inside your app: no tool loop, no usage metering per user and no plan caps.

Side by side

ShipFastAgentwright Starter (ours)
PriceOne-time. Starter $199 (list $299), All-in $249 (list $349), bundle with the CodeFast course $299 (list $648). Unlimited projects. Lifetime updates are on the All-in card and greyed out on the Starter card.$99 at the launch price for the first 50 licenses (code LAUNCH is applied at checkout), then $149. One-time payment. Source code for one person or one team, unlimited projects, every future version included.
RefundNo refund once you have access to the repo (FAQ).14 days, no reason needed.
Built forA general SaaS or web app: payments, database, login, emails, SEO and blog, UI components.A Claude agent product billed per user, built by one developer or a small team who will make their own UI.
Model calls inside your appNo in-app LLM integration listed. The FAQ covers AI code editors, which help you write the app.Claude only, through the Anthropic TypeScript SDK. Its own streaming tool loop with parallel tool calls, a 12-turn limit and prompt caching. MCP servers can be connected as tools.
Usage metering per userNot mentioned on their page.Yes. One usage row per model turn with input, output, cache read and cache write tokens, the cost and the id of the price row used. A monthly token cap per plan, checked before each model turn. A plan with overage_allowed set (the seeded Team plan has it) keeps running past the cap; that usage is recorded, and the kit does not bill for it.
Billing providersStripe or Lemon Squeezy.Lemon Squeezy or Polar, chosen with one setting (plan subscriptions and one-time license orders through signed webhooks). No Stripe.
DatabaseMongoDB or Supabase.Postgres with Drizzle ORM. Any Postgres should work; the deploy guide uses Neon.
Sign-inGoogle OAuth and magic links.GitHub sign-in, plus optional email magic links through Resend (Auth.js v5).
Landing page, blog, SEOIncluded: sections to copy for a landing page, a blog, SEO tags and UI components.Not included. The home page is a stub, and there is no blog or component library.
Language and routerJavaScript or TypeScript; app router or pages router.TypeScript, Next.js 16 app router.
TeamsNot mentioned on their page.No. Each user has one plan.
Evals and testsNot mentioned on their page.An eval runner with golden cases, 64 unit and database tests, and a GitHub Actions workflow.
Track recordThe page says 8,447 makers use it. The FAQ says the premium plan comes with a Discord of 5,000+ makers.New. The first release was on 2026-11-01. There are no customers yet, so there are no reviews.

On a phone, scroll the table sideways. "Not mentioned on their page" means we did not find it in the sources below, not that the product lacks it.

Sources for ShipFast:

  • shipfa.st (home page with pricing and FAQ), checked November 12, 2026.

Agentwright Starter: the kit's source, version 0.5.1. What's inside lists every file, and the refund and license terms are public.

What an agent product needs that a general kit leaves to you

These parts are specific to a product where a model does the work and every request costs you money. Each one links to a post with the code, so you can build it yourself on ShipFast or on any other kit. The file names show where the same part lives in Agentwright Starter.

A tool loop with a turn limit, streamed to the browser

In the kit: lib/agent/loop.ts · app/api/agent/route.ts

The agent calls the model, runs the tools it asks for, sends the results back, and repeats until the model is done. The loop needs a maximum number of turns, has to return all tool results of a turn in one message, and has to stream text to the browser while it works.

The code, explained: Stream a tool-using Claude agent to the browser with SSE in Next.js · Claude Agent SDK vs rolling your own tool loop

A usage row for every model call, per user

In the kit: lib/billing/metering.ts · lib/billing/pricing.ts

Every response reports input, output, cache read and cache write tokens. To know what a customer costs you, store those per user together with the cost in dollars and the price that was used, because model prices change.

The code, explained: How to meter Claude API costs per user in a Next.js SaaS · Model prices change. Snapshot them on every usage row

Plan caps and rate limits

In the kit: lib/billing/metering.ts · lib/ratelimit/index.ts

A subscription price is fixed and usage is not. Each plan needs a monthly cap that is checked before the model is called, and a requests-per-minute limit that answers 429 instead of spending tokens.

The code, explained: Rate limiting an LLM endpoint with only Postgres · How to price an AI SaaS plan so heavy users don't eat your margin

Prompt caching that stays on

In the kit: lib/agent/loop.ts · lib/agent/system-prompt.ts

An agent sends the same system prompt and tool list on every turn. Caching them lowers the input cost, and small changes can switch the cache off without any error.

The code, explained: Prompt caching in a Claude agent loop: where the breakpoints go and what silently breaks them

Evals you can run in CI

In the kit: evals/run.ts · evals/cases.ts

Changing a prompt or a tool description changes what the agent does. A small set of golden cases, run through the real loop, shows a regression before your users see it.

The code, explained: Evals for a Claude agent in 140 lines of TypeScript

Tools from MCP servers

In the kit: lib/agent/mcp.ts

If the agent should use tools that already exist as MCP servers, the loop needs a client that connects to them and registers their tools next to your own.

The code, explained: Use MCP servers as tools in your own Claude agent loop

Two ways to get there

ShipFast, plus these parts written by you

This fits if you already own ShipFast, or if the landing page, the blog, Stripe billing and the MongoDB or Supabase setup are what you want most. You add the six parts above yourself. The linked posts have the code and do not need our kit. Leave the most time for metering and plan caps, because a mistake there shows up on your API bill.

A kit built around the agent, plus a public site written by you

Agentwright Starter has the six parts wired together, with tests. It has no landing page, blog or component library, and its billing code is for Lemon Squeezy and Polar, not Stripe. You would build the public site yourself and, if you need Stripe, your own billing code.

Stay with ShipFast if…

  • Your product is a general SaaS, and AI is a small part of it or no part.
  • You want a landing page, a blog, SEO tags and UI components on the first day. Agentwright Starter has none of them.
  • You want Stripe or MongoDB. Agentwright Starter is Postgres, with Lemon Squeezy or Polar for billing.
  • A large community matters to you. ShipFast's page says 8,447 makers use it; Agentwright Starter is new and has no customers yet.

Pick Agentwright Starter if…

  • The product is a Claude agent, and your customers pay for a plan.
  • You need to see what each user's model calls cost, and to stop a plan at its cap.
  • You will build your own UI and public site anyway.
  • You want to read code before you pay and be able to change your mind: What's inside shows three files in full, and the refund window is 14 days.

Where Agentwright Starter is weaker

The same list as on our nine-kit comparison:

  • It is new. The first release was on 2026-11-01 and the current version is 0.5.1. There are no customers yet, so there are no reviews and no showcase of apps built with it.
  • Claude only. The agent loop, metering and evals are built on the Anthropic SDK. If you need OpenAI or Gemini too, another kit fits better.
  • No Stripe. Your app's billing is Lemon Squeezy or Polar. The Polar adapter is new in version 0.5.0: it has 49 automated tests, and on 2026-11-13 we ran a subscription through it in Polar's sandbox (Polar's real servers, a test card, no real money): checkout, plan change, cancellation and revocation each set the user's plan as documented. A renewal and a failed renewal come a month after a purchase, so those are covered by tests with Polar's sample events and not by a live run, and no real-money payment has gone through it. Its docs list the same five steps to check with your own Polar account before you take payments.
  • No teams, organizations or seats. Each user has one plan; "team" is only a plan name.
  • Minimal UI: sign-in, chat, usage, pricing and an admin page for licenses. The home page is a stub. There is no marketing site, blog or component library.
  • No live demo. Instead, /inside/ lists every file and shows three of them in full.
  • No RAG or search, no file uploads, no job queue. The plan cap is a monthly token count, and the kit does not charge for usage past it (on a plan with overage_allowed set, such as the seeded Team plan, usage simply continues).

Questions

Can I build an AI app with ShipFast?

Yes. Its page says makers have built AI tools with it, and you can call any model API from a ShipFast app. What the page does not list is the code that runs a model inside your product: the tool loop, usage metering per user and plan caps. Those would be yours to write.

Is Agentwright Starter a drop-in replacement for ShipFast?

No. It has no landing page, blog, SEO setup or component library, and it bills through Lemon Squeezy or Polar, not Stripe. It covers the part that ShipFast's page does not list: the agent loop, metering, plan caps, MCP tools and evals.

Which one costs less?

On November 12, 2026 the two pages showed ShipFast Starter at $199 and Agentwright Starter at $99 for the first 50 licenses, then $149. Both are one-time payments. They cover different work, so compare what you would still have to build with each.

Can I use both together?

They are separate codebases with different databases and billing code, so you would be merging two apps. It is simpler to pick one as the base and build the missing parts from the posts linked above.

More comparisons

Checked on November 12, 2026. We re-check every month. If a fact about your product is wrong or out of date, email hello@agentwright.app and we will fix it.