.Code

Engineering insights for your code and the team behind it. .Code reads your GitHub and GitLab organizations, every repository's whole history, and says in plain words how the work is progressing, how maintainable the code is, who it rests on, and how the code AI writes holds up.

On a waiting list for now. Accounts open by invitation: join the list, and we email you when yours is ready.

The .Code insights page of the pallets organization: 11 repositories, 2,012 contributors and 21,244 commits, and an executive summary saying more work is shipping but more of it is in code only one or two people know

Questions commit counts can't answer

Who moves the product forward?

Scores and badges on every commit tell features and fixes apart from churn, for every person and repository.

Who knows this code?

Ownership and truck factor show where knowledge sits with one person, before that becomes a problem.

Where does the effort go?

Hotspots and period-over-period shifts point at the files and areas taking the work, and the ones cooling down.

How much of our code comes from AI?

AI's share by person, repository and tool — as an engineer's sidekick, or as an agent on its own.

Contribution insights that go beyond raw activity

One engineer's page in .Code: 104 commits in the last 3 months against 57 before, additions and deletions, working days in a row, the languages they work in, the share of their commits in each value tier and badge, and a calendar of the last year
Every contribution, scored
A value score per commit — from code-line yield, file composition, message quality and the branch it lands on — mapped to six readable tiers: Stellar, High, Solid, Modest, Trivial, Weak.
Engineering personas
The badges on each person's recent commits suggest a working style — or generalist, when none leads: Builder, Maintainer, Steward, Refactorer, Platform, Maverick.
Contribution badges
Every commit is classified by its message, the files it touches and its yield: feature, tests, documentation, vendored files.
Calendar and streaks
A working-day heatmap of each person's year, their current streak, and a feed of every commit with its score and badges.
Smart auto-ignore
Inactive and empty repositories drop out of every total automatically — restore any with one click.

How each codebase changes, and who changes it

The year behind, the last three months and where it is heading, then a table of figures — work shipped, people contributing, maintainability, tests, bug fixes and AI — each better, worse or holding
AI, as sidekick and as agent
Commits are attributed to a person alone, a person with an AI sidekick, or an AI agent on its own — read from co-author lines, tool signatures and the agents' own accounts. See who works with AI, which agents commit for themselves, and AI's share across repositories and code.
Hotspots
The files that change most, weighted by their size — where the effort goes, and who spends it. The top 25 per repository.
Ownership and truck factor
Who knows which files, and how many people would have to leave before most of the active code had nobody who knows it.
Change coupling
Files that keep changing together — a file and its test, or a hidden dependency across modules.
Shifts over time
The last month, 3 or 6 months, or the year against the same length of time before — commits, people, test share and AI share — and the areas gaining or losing attention.

From your repositories to insight in three steps

  1. 1Make a profile

    A profile is one team's space: its sources, members and data. Many teams can share one .Code.

  2. 2Add organizations and repositories

    Whole GitHub organizations, GitLab groups — on gitlab.com or your company's own server — or single repositories, each read with one of the profile's read-only tokens, encrypted and never shown again.

  3. 3Sync, then explore

    A sync reads the code host once, so exploring never runs into its rate limits. Each repository is cloned into memory, measured and gone when the work ends — never on a disk, never in a log.

A demo profile's sources in .Code: the acme organization and the octo-labs/cli repository, each read from GitHub with the profile's key

Real engineering signal, not activity counts.