Skip to content
The Fabrica

Comparisons

The Fabrica vs Hype Stack

A composable TypeScript application platform that reaches five surfaces, against an opinionated Python foundation that goes deep on one.

Last reviewed · from Hype Stack’s own site and documentation

The short answer

Choose Hype Stack if

your team is TypeScript end to end, or you need mobile, desktop or an extension alongside the web app. Start free on the MIT core and add the packs you want.

Choose The Fabrica if

your backend work is Python, long jobs need their own worker process, and you are selling metered usage rather than seats.

At a glance

The dimensions a decision turns on

Twelve rows, not forty. Every answer about their product is numbered and links to the page it came from, so you can check it rather than trust it.

Backend
The FabricaFastAPI and async SQLAlchemy 2.0, in Python.3
Hype StackA Hono API server on Node.js, with Prisma owning the schema and migrations and Kysely owning the harder queries.3
Frontend
The FabricaNext.js and React, with Tailwind and next-intl. One web surface.3
Hype StackReact 19 with TanStack Router, Tailwind and shadcn/ui — and the same codebase reaches an admin app, an Expo mobile app, an Electron desktop window and a browser extension.2
Data layer
The FabricaPostgreSQL with Alembic migrations. Redis is the Celery broker.3
Hype StackPostgreSQL, with Valkey (Redis) used for caching and session storage.2
Source model
The FabricaA commercial source licence. You get the repository and own what you build in it; it is not open source.1
Hype StackAn MIT-licensed core template and CLI, with features sold as packs. Packs are copied into your repository, and the code you pull stays yours whether or not you renew.1
Pricing
The FabricaSolo $379, Team $749. One payment, not a subscription.1
Hype StackThe core is $0 and MIT. Premium is either a one-time purchase of an individual pack, which never renews, or an All-Access licence billed yearly.1
Payments
The FabricaPaddle, acting as merchant of record, so sales tax is filed by the provider rather than by you.8
Hype StackBilling arrives as a pack, and you pick a provider: Stripe, Lemon Squeezy or Polar.4
Background work
The FabricaCelery workers on Redis, deployed as their own process, with scheduled beat tasks and a reconciler that repairs missed state.5
Hype Stack“In the template as an in-process scheduler. No external queue; jobs are stateless sweepers over your own tables.” The scheduler is croner, running inside the API, and each fire takes a Postgres advisory lock — so there is nothing extra to deploy.2
Usage and credits
The FabricaAn append-only credit ledger: every movement is one INSERT and the balance is derived, so concurrent spends and refunds reconcile.4
Hype StackThe AI-chat pack lists usage caps among the things it ships.4
Coding-agent context
The FabricaOne canonical instruction file at the repository root, reachable from AGENTS.md and a Cursor rule, plus 49 decision recipes an agent can read instead of guessing.7
Hype StackEditor rules for Cursor, T3 Code, Claude, Codex, Copilot, OpenCode and Windsurf, a starter set of agent skills, and an MCP server.5
Extension model
The FabricaReserved _domain/ directories: upstream does not overwrite the files you create there, which shrinks the surface an update can collide with rather than removing it.6
Hype Stack“Packs are copied, not installed, the same way shadcn/ui works.” The CLI merges the Prisma schema and offers to run the migration.4
Deployment assumptions
The FabricaDocker Compose locally. Postgres, Redis and a separate worker process are assumed to exist wherever you deploy.5
Hype StackPostgres, Valkey and object storage in Docker locally, or on Fly or Railway. Every app deploys independently and talks to the API over HTTP.2
What you can read first
The FabricaPublished engineering documentation, a public inspection repository, and a walkthrough recorded from a release tag.9
Hype StackThe core template and CLI are public on GitHub under MIT, alongside public documentation, a pack catalogue and a template gallery.6

Philosophy

What each one is optimising for

Feature tables compare what two products contain. This is the part that decides whether either fits: what each one decided to be good at, and what it gave up to be good at it.

A platform that composes

Hype Stack optimises for reach and for choice. The free core gives you five deployment surfaces from one TypeScript codebase — web, admin, mobile, desktop, extension — and the features you actually need arrive as packs you copy in, the way shadcn/ui components do. Auth, billing, teams, notifications and AI chat are catalogue entries, not fixed parts of the foundation.

That is a real architectural position, not a shortcut. Copying code instead of installing a dependency means every line is in your repository and readable. Picking your billing provider at compose time means the foundation never has to be right about which one you will want. And keeping jobs in the API process means there is no queue to run: their own docs say so plainly, and “nothing extra to deploy” is a genuine operational saving on the day you launch.

The cost of composability is that integration is yours. Two packs are two decisions; the seam between them is a third, and it is not decided until you compose them.

A foundation that has already decided

The Fabrica optimises for the opposite thing: fewer surfaces, more pre-made decisions, and the decisions written down. There is one web surface and one Python backend, and the parts that touch money, jobs and personal data are integrated rather than optional — the credit ledger, the webhook idempotency tables, the account-erasure path and the audit log were built against each other and are tested together.

The bet is that on an AI-heavy product the expensive part is no longer typing the code. It is knowing which decisions are load-bearing, and an agent asked to add a feature will happily invent a credits_remaining counter unless something in the repository explains why the ledger is append-only. So that reasoning ships with the code and an agent reads it.

The cost is symmetrical and worth saying: you inherit our choices. Python on the backend, Paddle at the checkout, Celery behind the jobs. If any of those is wrong for you, the composable model is the better fit and no amount of documentation changes that.

Fit

Who should choose which

Choose Hype Stack

Hype Stack is the stronger choice, and it is not close, when:

  • Your team is TypeScript end to end and putting Python in the stack means hiring for it or context-switching all day.
  • You need more than a web app — a mobile app, a desktop build or a browser extension from the same codebase is a large amount of work you would otherwise repeat.
  • You want to start free. The core is MIT and costs nothing, which is a materially different risk profile from paying before you have read the code.
  • You would rather assemble than inherit: pick the billing provider, skip the packs you do not want, and keep the foundation thin.
  • Not having a queue or a worker to operate is worth more to you than isolating long jobs from the request path.

Choose The Fabrica

The Fabrica is the better fit when:

  • Your backend is Python because the work is — model calls, document pipelines, data processing — and a TypeScript API means a second service or a rewrite.
  • Jobs are minutes long, not seconds, and you want them off the request path in their own worker process from the first deploy.
  • You sell usage rather than seats, and you want a ledger that survives concurrent spends and refunds rather than a counter you will have to replace later.
  • You would rather read why a decision was made than re-decide it — and you want your coding agent to read the same thing.
  • Selling into the EU as a solo founder makes a merchant of record worth more than provider choice.

Check it yourself

Do not take our word for our own column

Everything claimed above about The Fabrica is published, and most of it predates this page.

Still deciding?

Read the code before you read the price. If it is not the right foundation for what you are building, Hype Stack is a good product and this page has tried to say why.

Pricing and licence →All comparisons

Sources

Where the Hype Stack facts come from

Read on . Where a row says a fact was not established, it means we could not confirm it from public material — not that the product is missing it.

  1. 1Hype Stack — plans and licences
  2. 2Hype Stack docs — technologies
  3. 3Hype Stack docs — backend overview
  4. 4Hype Stack — feature packs
  5. 5Hype Stack docs — working with AI
  6. 6BetterTyped/hype-stack on GitHub

Where the The Fabrica facts come from

Read on .

  1. 1Pricing
  2. 2Under the hood
  3. 3Architecture
  4. 4Credits are a ledger, not a counter
  5. 5Background work
  6. 6The extension boundary
  7. 7What a coding agent is handed
  8. 8Why Paddle, not Stripe
  9. 9Public inspection repository