Paste a file Git left with conflict markers and see both sides of every conflict, the ancestor where one is recorded, and any marker left without a partner.
Waiting for a file with conflict markers
Paste the file Git left behind. Nothing is uploaded and nothing is written to your repository.
Each conflict appears here
Both sides of every conflict, the ancestor where the file records one, and any marker left without a partner.
Merge Conflict Inspector runs entirely on your device — nothing is uploaded.
You’re on the Free plan. Compare plans
A merge conflict in a file you did not write is hard to read. Seven angle brackets, a label that is a branch name or a commit hash, and two blocks of code with no indication of which one is yours. This separates them: every conflict listed with both sides, the lines they occupy, and the labels Git wrote.
Taking a side is a bulk edit and the page calls it that. Choosing ours for every conflict is right when you know one branch is correct throughout and wrong the moment the answer differs per conflict, which is most of the time. Nothing here merges anything intelligently, and a tool that implied otherwise would be lying about the only thing that matters.
Two problems it catches that reading the file does not. The first is a conflict whose two sides are byte-identical: a conflict over nothing, usually a line-ending or trailing-whitespace difference Git could not reconcile on its own. It is flagged, because resolving it either way gives the same result and the time spent reading it is wasted.
The second is an orphan marker. A closing marker with no opening line, or a separator sitting outside any conflict, is what a half-finished manual resolve leaves behind. It is easy to miss by eye, it compiles in several languages, and it is the single best reason to run a file through this before committing it.
Conflicts written in diff3 or zdiff3 style carry the common ancestor as well as both sides, and where that is present it is shown. The default two-sided style does not record it, so the tool says the ancestor is unavailable rather than inventing one. Nothing is uploaded and nothing is written to your repository: this reads text you paste and hands back text you copy.
No. It takes one side for every conflict in the file, which is a bulk edit rather than a merge. When the right answer differs from one conflict to the next, read both sides here and edit the file yourself.
Ours is the first block, between the opening marker and the separator — the branch you had checked out. Theirs is the second. During a rebase those are swapped relative to what most people expect, which is why the labels Git wrote are shown beside each side.
Because they are, byte for byte. Git can still produce a conflict when the surrounding context differs, or over line endings and trailing whitespace. Resolving it either way gives the same result.
A closing marker or separator with no opening line, left behind when part of a conflict was deleted by hand. It often compiles, so it survives to a commit and then confuses everyone who reads the file afterwards.
A Git setting that writes the common ancestor into the conflict as well as both sides, so you can see what each branch changed rather than only what each ended up with. Where a file has it, this shows it.
No. It cannot see your repository at all. You paste text in and copy text out, and both stay in your browser — which is what makes it safe for a file that still has credentials in it.