Keeping the code in your own repository

What it is for

Some people want the code somewhere they control, or want a developer to work alongside the builder. If you ever leave, the code stays with you.

Changes made here go there, and changes made there come back.

What you need

A GitHub account and either a sign-in with GitHub or a personal access token. Either one behaves the same.

A repository and a branch. The branch name is checked before anything runs, and a private repository is fine.

The order to do it in

Connect the account, choose the repository and branch, then send the current files across once.

An empty repository is easiest to start with. Existing code turns the first screen into a conflict.

After that, keep changes on one side at a time. Two sides at once is what makes a conflict.

How to check it worked

The connection shows one of five states: in sync, ahead, behind, conflict, disconnected.

Behind means the repository is ahead. Pull first, then carry on editing.

A conflict names its reason — local changes, content conflict, the remote was rewritten, the head moved, the branch was deleted, or the credential is gone.

What it will not do

• Your token is never written into the project record. It stays in a vault in memory.
• A conflict is not resolved for you. You are told which files disagree.
• The repository is yours. Removing the connection does not remove the repository.
• Only code travels. What your visitors left in the database does not go to the repository.