software-atlas

Fleet-Wide Workspace

Fleet-wide line-of-code counter: per-server agents report first-party LOC

Fleet-wide line-of-code counter: per-server agents report first-party LOC (heuristics plus .locignore overrides) and the service stores and visualizes it in a treemap.

Fleet-wide Line-of-code Counter

Fleet-wide line-of-code counter: per-server agents report first-party LOC (heuristics plus .locignore overrides) and the service stores and...

Core Workflow

Designed around fleet-wide workflows with secure authentication and API-first boundaries.

Operations Ready

Includes health checks, versioning, and deployment-ready infrastructure from day one.

Fleet LOC

How these counts are produced

finance / bubble: an agent runs scc by file over each app directory, excludes known dependency and data directories, then classifies vendored or generated files using shared path/content heuristics plus any repository .locignore.

github: a repository enters the dataset when an authenticated client requests an owner/repo measurement, including dependency surveys and the former GitTrends workflow. Software Atlas shallow-clones the repository’s current default-branch HEAD, runs scc by file, applies generic vendored/generated heuristics, and stores the snapshot.

“First-party” is a heuristic classification, not a guarantee. GitHub clones cannot use repository-specific .locignore rules, so unconventional vendor trees, generated tables, fixtures, and test data can inflate large repositories. Counts represent snapshots taken at different times.

Loading…