You removed the build output and deleted an old archive, but the repository is still huge. Often the missing piece is Git's own history.
Deleting a file does not remove every version of it
The working tree contains the version you have checked out today. The hidden .git directory contains objects, pack files, references, and other data Git needs to reconstruct earlier commits and branches.
If somebody committed a database dump, installer, video, or generated binary six months ago, deleting it in a later commit removes it from the current tree. The old content may still be part of repository history, which means the clone can remain much larger than the visible project.
Find out whether the space is in the project or in .git
Scan the drive and look for Git repositories in SizeTrend's Overview pane. SizeTrend identifies repository roots and lets you inspect the storage beneath them. It can also list linked Git worktrees, which are easy to forget after a branch or experiment is finished.
Open the repository location and follow the largest folders. If .git\objects\pack accounts for most of the size, you are looking at packed Git history. If the working tree is larger, build artifacts, dependencies, test results, or ignored files may be the real issue.
Use Git tools before touching Git data
Run git count-objects -vH in the repository for a Git-aware summary. If you suspect one large object in history, use your normal repository maintenance tools to identify it and decide whether it belongs there.
Do not delete pack files or random folders inside .git. Rewriting published history to remove an old object affects everyone who uses the repository and needs to be coordinated. A normal Git cleanup may discard unreachable objects, but it will not erase content that reachable commits still need.
Old clones and linked worktrees are a different case
If an entire local clone is no longer needed, SizeTrend's repository-specific Recycle Bin command performs a local Git preflight before offering to move it. Check for uncommitted or untracked work first, and make sure anything important exists somewhere else.
A linked worktree should be removed with git worktree remove, not by deleting its directory. Git keeps bookkeeping for worktrees in the main repository, so a manual folder deletion can leave a confusing half-removed entry behind.
See the whole repository, including the hidden parts.
SizeTrend can separate a large working tree from the Git data and old worktrees around it.