Know which files matter

Start by separating the files you prepare from the files Divekit generates later. That distinction prevents a lot of beginner confusion.

At the beginning of the course, you do not need to think about every internal Divekit file.

What matters first is simpler:

• there are setup files
• there are generated state files

Later in the course, the relevant part of the project will look roughly like this:

vehicle-workshop/
└── .divekit/
    ├── individualization.json
    └── distributions/
        └── M0/
            ├── config.json
            ├── individuals.json
            └── remotes.json

Setup files

These files are the minimum you need to care about in the beginner workflow:

• config.json describes one rollout
• individualization.json defines what may vary across the distributed repositories

In the normal beginner flow:

• config.json is created by Divekit commands such as divekit init
• individualization.json is the file you will edit manually

Generated state files

These files are created by Divekit during distribution:

• individuals.json records which concrete variation was selected for each UUID
• remotes.json records which remote repositories were created

These are result files.

You do not prepare them by hand.

Member input is stored separately: Divekit saves it as a timestamped file in your user-level Divekit directory (~/.divekit/members/) — outside the project — and references it from config.json.

Why the distinction matters

A beginner mistake is to treat every JSON file as if it should be prepared manually.

The safe rule is:

• prepare the actual input files
• let Divekit generate the result files
• inspect generated state instead of editing it

Check

Before moving on, make sure you can answer these questions: