1 min to read

Git Without Git: Building a Repo Manually

Did you know we can create a Git repository without ever running git init?

Git is, at its core, just a structured directory. With a few simple commands, we can manually build the essential .git/ structure yourself. This is a great way to demystify Git internals and understand what happens behind the scenes.

🧱 Step 1: Create Your Project Directory

mkdir my-repo
cd my-repo

🗂️ Step 2: Add the .git Directory

mkdir .git

This folder will store all of Git’s internal data.

📦 Step 3: Add the objects Directory

mkdir .git/objects

This is where Git stores all content (blobs, commits, trees) in hashed object form.

🌿 Step 4: Add the refs Directory

mkdir -p .git/refs/heads

Git uses refs/ to manage branches and tags. refs/heads/ contains pointers to local branches.

🧭 Step 5: Create the HEAD File

echo "ref: refs/heads/master" > .git/HEAD

The HEAD file tells Git which branch is currently checked out. Here, it’s set to master.

✅ Step 6: Verify with Git

git status

We should see:

On branch master

No commits yet

nothing to commit (create/copy files and use "git add" to track)

Congrats—we just built the foundation of a Git repository without running git init.

📝 Why Try This?

Manually creating a Git repository is a fun way for us to learn about Git’s internals. It’s not something we’ll do every day, but it’s a great exercise for understanding how Git works under the hood.