The Boring Two Days of a New Project, Already Done
You can write all of this. That is not the problem. Fabricate generates the auth, the schema, the API routes and the deploy config so the project starts at the part that is actually yours. Standard React and TypeScript, downloadable on a paid plan.
None of this is difficult. All of it is required, and you have written it before.
Auth, database client, migrations, API routes, deploy config. You have written it before and you will write it identically again.
The idea was good. It died in the scaffolding, three weekends in, before the interesting part started.
The client means one form. Delivering it properly still means auth, storage, validation, hosting and a domain.
Frontend, backend, schema, infra. Each context switch has a warm-up cost, and a solo project charges you all four.
Generated, configured and deployable, so the first commit you care about is a feature.
Auth, schema, API routes and deployment generated and wired together. Open the project at the feature that made you want to build it.
React, TypeScript, Tailwind, Hono, Drizzle, Cloudflare Workers and D1. No framework to learn and nothing proprietary to unpick.
On a paid plan, take the whole codebase into your own editor and carry on. The platform is a starting point, not a host you are stuck with.
Ask for a feature in plain English, then read the diff like any other pull request. You stay the one deciding what is correct.
Ship to Cloudflare's edge without writing pipeline config. HTTPS and global distribution are already handled.
Typed queries against a real database, so the data layer is something you can extend rather than something you replace.
Reach the part that interested you while the interest is still there.
Scaffold in an afternoon and spend the billable hours on what the client is actually paying to get.
Put a working version in front of a stakeholder before committing a sprint to the real one.
The admin panel nobody wants to staff, built well enough that the team stops asking you to run queries.
Spend the weekend on the idea rather than on auth and deployment.
See how a complete app gets assembled end to end, then take the patterns you like.
The estimate that ruins a side project is rarely the feature. It is everything around it. Wiring authentication, choosing and configuring a database client, writing the first migration, setting up the API layer, getting deployment and a domain working. None of it is difficult and all of it is required, which is why a two-hour idea reliably becomes a three-weekend commitment that ends before the interesting part.
What is worth generating: The parts that look nearly identical across every project you have built. Auth, schema and migrations, CRUD endpoints, deployment configuration. There is no craft being lost here, and no advantage in writing them again by hand.
What is not worth generating: The logic that makes the product yours. Generation gives you a competent first draft to review, not a decision about how your domain should work. Read the output the way you would read a colleague's pull request.
Leaving when you want to: The output is standard React and TypeScript on Cloudflare infrastructure. On a paid plan you can download it or push it to GitHub, so nothing about starting here commits you to staying.
Generate the auth, schema, API routes and deploy config, then take the code into your own editor. Start free.