Find review hotspots in unfamiliar Git histories.
The demo is captured from the real CLI using synthetic data. Transcript · regenerate with python3 scripts/demo.py.
An unfamiliar repository has years of context hidden in Git. GitLore summarizes churn, a lightweight complexity estimate, contributors and activity so you can choose where to start reading.
| Start with | Good for | Where GitLore helps |
|---|---|---|
git log |
Read commits in sequence | GitLore aggregates file hotspots and activity. |
git shortlog |
Count contributors | GitLore combines contributors with churn and review signals. |
git clone https://github.com/LucasXlab/gitlore.git && cd gitlore
python3 -m pip install .
python3 scripts/demo.pyRequires Git and Python 3.10+. The install includes Rich and Textual. Optional AI support requires python3 -m pip install ".[ai]"; core analysis needs no API key. Open the generated HTML example.
Combine churn and a heuristic complexity estimate. Scores prioritize attention; they are not defect probabilities.
--diff main...HEAD weights changed files by historical risk to suggest a reading order.
Export standalone HTML or JSON, or explore the repository in the Textual TUI.
--explain is optional and uses your configured API endpoint. It sends a commit patch only when explicitly selected.
gitlore --repo /path/to/project
gitlore --since 2026-01-01 --top 10
gitlore --diff main...HEAD
gitlore --html report.html
gitlore --json
gitlore --tuiflowchart LR
A[Local Git history] --> B[Churn / contributors / timeline]
C[Tracked source files] --> D[Complexity estimate]
B --> E[Hotspot and review ranking]
D --> E
E --> F[Terminal / TUI / HTML / JSON]
G[Explicit commit patch] -. --explain .-> H[Configured API endpoint]
Core analysis stays local. HTML/JSON reports include repository paths and contributor names and must be reviewed before sharing. --explain sends up to 12,000 patch characters to the configured endpoint; file contents may contain sensitive data. Installing dependencies can access package registries. See PRIVACY.md for data boundaries.
python3 -m pip install -e ".[dev]"
python3 -m pytest -qThe demo uses a temporary fixture and does not inspect your personal repositories or agent homes.
- Improve language-aware complexity estimates.
- Add interactive per-file history drill-down.
Read CONTRIBUTING.md. Report reproducible problems in Issues. Follow SECURITY.md for sensitive reports.
MIT.