How Git Works Internally: Objects, Branches & the .git Directory

Git is a content-addressable filesystem — a key-value store where the SHA-1 hash of content becomes its identifier. The entire database lives in the hidden .git directory, storing everything as four object types: blobs (file content), trees (directory structure), commits (snapshots), and tags (named markers).

When you git add, file content is hashed into a blob and tracked in the staging index. When you git commit, the staging area becomes a tree object, which gets wrapped with metadata into a commit object, and the branch ref updates to point to it. A branch is literally a 41-byte text file containing a commit SHA — creating one is just writing a file. Pack files with delta compression keep repositories efficient despite storing full snapshots.

Understanding this model demystifies merge conflicts, detached HEAD states, and makes recovering lost commits trivial instead of terrifying.