← the index
marktext/marktext ↗
81.7%
of 536 commits were committed by an AI agent — a floor, not an estimate
81.7% agent · top contributor 91.1%12 contributors
individual

One person is driving this. The headline number reflects a single contributor's workflow, not a project-wide shift.

438
agent commits
98
human commits
12
agent contributors
91.1%
largest contributor

Assisted or autonomous

438assisted
0autonomous

Assisted — a developer’s identity authored the commit and the agent added itself as co-author. The agent ran git commit during a working session, with a person there. Autonomous — a bot identity is the author, no human in the loop. Across the whole index, 93.7% of agent commits are assisted.

Agent share by month

Mar26
May26
Jun26
Jul26
Sep26

Peak month: 90%. Bars are scaled to that peak, not to 100%.

Which agents

Claude CodeAnthropic43380.8%
GitHub CopilotGitHub71.3%

Verify it yourself

Sample commits that matched. Every one links to GitHub — check the trailers.

  • 0e156b0afix(muya): remove the selected rest of a list or quote when a cut starts iClaude Code
  • e5422625fix(muya): keep the code block when a cut ends in its language line (#5371Claude Code
  • 555f3a72fix(muya): collapse only the nested code block when a cut starts in its laClaude Code
  • 26825e77fix: replace remaining absolute raw.githubusercontent.com URLs with relatiGitHub Copilot
  • f7100c27fix: build error when build mac target (#4232)GitHub Copilot
  • 0c8eadb6i18n: translate sponsor section in all language READMEs (#4209)GitHub Copilot

What was excluded

Commits in window543
Merge commits removed−7
CI / dependency bots removed−0
Considered536

Bots are excluded from the denominator entirely. Counting dependabot as “AI-written code” is the fastest way to produce a meaningless number.

How this was measured

Measured over February–August 2026, cloned directly. Detection is deterministic string matching against signatures agents write about themselves — no heuristics, no style analysis. Merge commits and CI bots are excluded from the denominator. The figure is a floor: squash merges strip trailers, so real usage is higher. Contributor counts are aggregate — no individual is named.

Run it yourself: npx qmmit-cli marktext/marktext

Scan another repository