Code, data, and exports

Getting your code and database out of Floot, and how to bring an existing project in.

Do I own the code?

Completely — code, data, and IP. See Security and data.

How do I download or export my code?

There is no export tool on the connector, so your assistant cannot produce the download for you. It can read any file in the project, so it can copy specific code into your chat, but the runnable zip has to come from the browser.

Open your project on floot.com, click the project name in the top bar, and choose Get Code, then Download as ZIP. The download includes instructions for running it, and comes with every paid plan. If you plan to host it yourself, read Self-hosting first.

Can I work on the project files on my computer?

Yes, on a paid plan, in Chrome or Edge: Get Code → Folder Sync connects a folder on your computer to the project. Floot compares the two and lists what would move each way — changes to send to Floot, changes to bring into your folder, ones it merged on its own, and conflicts for you to decide — and nothing moves until you review the merge and apply it. It syncs the project's own files, so it is not a way to import an unrelated codebase.

The connection to the folder is remembered by your browser, so clearing site data disconnects it: connect the folder again from Get Code. The server.ts and README that let you run the app elsewhere are not part of a synced folder; they come only with the zip.

Can I connect my project to GitHub?

Not directly: Floot does not sync with a Git repository. On a paid plan you can keep a folder on your computer in sync with the project (Get Code → Folder Sync) and commit from that folder, or download the zip and push it yourself.

Can I import an existing codebase into Floot?

Floot cannot run your existing codebase as-is — it rebuilds it. What you get is a Floot-native app that looks and behaves like the original, not a port, so third-party API keys and data migrations still have to be redone by hand.

There is no import tool on the connector — nothing takes a repository or an archive and turns it into a Floot project. Handing your assistant a zip does not import anything.

What works instead: if your assistant can already read your source — Claude Code, Codex, and Cursor all run on your machine — point it at the repository and ask it to rebuild the app in Floot page by page. It writes the pages, endpoints and schema with the same tools it would use on a new project. Give it the schema and one page at a time rather than the whole repo at once.

To bring part of one Floot project into another, give your assistant both project links and say what to copy. It can read the files of any project on the account it is connected to.

Can I export my database?

Yes, but it is a browser step — there is no connector tool that produces a database dump.

Open your project on floot.com and go to the Services tab, then Data.
Select your Floot database and open a table.
Click the cog icon in the table header to open Database Settings, and under Export click Download Database. The export runs in the background, however large the database is; Download appears when it is ready and works for 24 hours. Tick Structure only first to get the schema without any rows (tables, indexes, constraints and functions), for documentation, an audit or planning a migration.

The file is a gzipped pg_dump (.sql.gz) without owners or grants, so it restores into any PostgreSQL database. It holds the database's schemas, tables, views, sequences, indexes, constraints, functions, triggers, row-level security policies and CREATE EXTENSION statements. It does not create database roles, and it leaves out object owners, grants and default privileges (GRANT, REVOKE, ALTER DEFAULT PRIVILEGES); a policy that names a role still names it, so that role must exist where you restore. Structure only also leaves out every row, including the current values of sequences. If the export fails, contact support with your project URL and we can generate one for you. Neither export nor import is available while your hosting balance is unpaid: from the moment it runs out, not only once your apps are taken down.

You can run up to 5 exports per database per day; one that fails does not count. Exports are manual. There is no scheduled export, so download one whenever you want a copy to keep. Unless you tick Structure only, the file contains the whole database.

Is my database backed up?

Floot keeps a 7-day recovery history of the built-in Floot database. On a paid plan you can put the database back to any moment in those 7 days, and undo it: Services → Data, open a table, then the cog icon → Backups & restore. Some older databases can't be restored from there; for those, contact support with the project link and the date and time you want to go back to, with your time zone. See Backups and restores.

Checkpoints restore your code, never your data. Before a risky change to your data, download an export first. More about how the database works is in Your database.

Can I import a database?

Yes, on paid plans, into the built-in Floot database, from the same dialog: cog icon → Import → Choose file. It takes a plain SQL dump (.sql or .sql.gz) or a pg_dump custom-format archive, up to 1 GB. Exporting is available on every plan.

An import replaces the whole database: every table and all data. You confirm by typing a word.
There is no undo in the dialog. If you need the earlier database back after an import, contact support.
It runs as one transaction: if the file fails anywhere, nothing is changed.
Your app waits on database calls while it runs, so pick a quiet moment for a large file.
Coming from another provider, export with pg_dump --no-owner --no-privileges so the file does not name database users that only exist there.