Creating & managing projects
Every project is a fullstack Next.js app with its own git repository, its own database, and its own production URL. This page covers the choices you make at create time and what you can change afterwards.
Creating a project
Click New project on your projects page. The first time you sign in, a welcome dialog offers the same thing.
- Enter a Name — up to 100 characters. It becomes the project title and the first version of your URL slug.
- Add a Description (optional) — up to 500 characters. It labels the project on your projects page; it is not passed to the AI.
- Pick a Scenario — the starter code the project begins with.
- Pick a Database.
- Click Create project.
Creating is not instant: the repository is seeded with the scaffold and the database is provisioned before the workspace opens.
The dialog also links to the community templates in the Showcase, if you would rather start from a finished project — see start from a Showcase template.
Starter scenarios
Four scaffolds. All of them are real, runnable code you can prompt against straight away.
| Scenario | What you get | Database |
|---|---|---|
| Blank Next.js | Minimal Next.js with Tailwind | Optional |
| Todo CRUD | Todo app with Prisma, server actions, and a working UI | Required |
| Landing Marketing | Hero, three features, a CTA — no server logic | None |
| Dashboard Authenticated | Sidebar, topbar, stat cards; sign-in is a stub | Required |
Picking Todo CRUD or Dashboard Authenticated selects PostgreSQL for you and disables No database — those scaffolds ship a schema and cannot run without one.
The scaffold is a starting point, not a cage. The AI reads what is actually in the repository on every run, so a project that started as a landing page can grow into anything.
Choosing a database
Three options: No database, PostgreSQL, MySQL. Picking an engine provisions a real database for that project, used by both your preview and your deploys.
The database choice is locked at create time. There is no way to add one later and no way to switch engines — if you need a different answer, create a new project.
The dialog marks this section (experimental). What you can do with the database once the project exists is covered in your project database.
Your projects page
Your projects page lists everything you own plus everything shared with you. Each card shows the name, the description, the URL slug, and when it was last edited. A project you collaborate on rather than own carries a Shared badge.
Three controls sit above the grid:
- All / Owned / Shared — filter by ownership.
- Search projects… — matches the name, the slug, and the description.
- Sort by Last edited, Name A–Z, or Oldest first.
When nothing matches, Clear filters puts the full list back.
Renaming a project
In the workspace header, open the avatar menu and choose Rename / edit to change the name and description. Editors can do this too — it is not owner-only.
Renaming also regenerates the URL slug from the new name. A deployment that is already live keeps serving its current URL; the new slug takes effect the next time you publish.
To set the slug directly, without touching the name, use Rename URL slug in the Deploy dialog. That path is owner-only, allowed once per 24 hours per project, and it takes the old URL offline immediately — details in publish & deploy.
Deleting a project
Delete project sits in the same avatar menu, and only the owner sees it.
Deleting is permanent and immediate. There is no trash and no restore window. Confirming removes the database, the git repository, every deployment and the containers behind it, uploaded attachments, build artifacts, and run logs. Credits already spent on its runs are not returned.
Export first if you want to keep the code — see export your source code.
Limits & notes
- Name up to 100 characters; description up to 500.
- Scenario and database are both fixed at create time.
- Your projects page loads the 50 most recently edited projects; search, filters, and sorting apply to that set.
- Renaming the name can change the URL slug. Renaming the slug directly is once per 24 hours and is refused while a deploy is running.
- Delete is owner-only, permanent, and not recoverable.
- Export, Download ZIP, and delete are owner-only. Editors can prompt, preview, and rename.
- Your plan sets how many projects you can have — 3 on Spark, 15 on Maker, unlimited above that.
- Limits may change during alpha.
Related
- Quickstart: idea to live URL — the whole path in six steps.
- The workspace — what each panel in a project does.
- Your project database — browsing data and connection strings.
- Publish & deploy — production URLs, history, rollback.