Changelog Publisher

Your customers hear what changed in words they can use, with anything that needs action first.

Every release contains a lot of work that matters enormously and means nothing to a customer: a refactor, a dependency, a test, an internal tool. Your Changelog Publisher takes all of it out, because a changelog that lists everything is one nobody reads — and then the single entry that needed somebody's attention gets missed.

It reads what actually shipped. From the merged work, not from a summary somebody wrote from memory. The two are never quite the same, and the difference is where the surprise lives.

Anything requiring action goes first. A setting to change, a migration, a deprecation with a date. Not buried under the new features, where the person who needed to act has already stopped reading.

A breaking change is called a breaking change. Not "improved consistency". Saying it plainly is the difference between a customer preparing for it and a customer's integration stopping on a Tuesday for no reason they can find.

Each entry says what it means, not what was built. "Exports now include tax as a separate column", not the name of a field in the code. Grouped by who is affected — everyone, administrators, developers — so a reader finds their part in one glance.

Security fixes name the area and stop there. Enough for a customer to know they should update, never enough to help anybody attack a version that has not.

And it catches the release nobody wrote up. Every week it compares what shipped against what was published and backfills the gaps, dated to when the change actually happened. Hotfixes are the ones most often missed, because they ship under pressure when the note feels optional — and they are exactly the changes a customer is most likely to notice on their own.