Most dev teams already have all their progress tracked, just not in a readable form.
Every commit, PR, and merge tells a story.
The problem is: Git doesn’t summarize it for you.
So teams still end up doing something surprisingly manual:
scrolling through commits
checking PRs
writing weekly updates by hand
We kept seeing this repeated across teams, so we built something to experiment with it.
What we built
We built Gitmore : https://gitmore.io/
It’s a tool that connects to your Git repositories and generates structured activity summaries.
Instead of manually writing updates, it produces:
what shipped
what changed this week
what’s still in progress
key repository activity highlights
It sits on top of GitHub/GitLab, it doesn’t replace them.
How it works (simple version)
Gitmore reads:
commit messages
pull request titles
merge activity
timestamps and repo structure
Then it groups and summarizes activity into readable updates.
No code access. No deep integration. Just Git metadata.
Why we built it
We noticed a recurring pattern:
Even in small teams:
devs spend time rewriting what Git already knows
reports are often rushed or skipped
context gets lost between engineering and non-engineering teams
The goal wasn’t to “add AI everywhere”,
it was to remove repetitive reporting work.
Example output
Instead of raw Git logs like:
fix login bug
update auth flow
refactor middleware
final changes v3
You get something like:
Authentication system updated
Login bug fixed affecting session handling
Middleware refactor improving request flow stability
Current limitations
This is still early, and there are real constraints:
commit message quality matters a lot
noisy repos produce noisy summaries
works best for teams with consistent Git practices
It’s not perfect, it’s an experiment in structuring dev activity better.
What we’re trying to learn
We’re mainly curious about:
Do teams actually want automated reporting?
Or is manual reporting still preferred for context?
Where does automation help vs where it becomes noise?
Feedback welcome
Especially interested in:
whether this fits into real dev workflows
what breaks in real-world repos
and whether this solves a real pain or just moves it around
Top comments (0)